Skip to main content

Full text of "Second West Coast Computer Faire Conference Proceedings"

See other formats



Jim C. Warren, Jr., Editor 

March 3-4-5, 1978 San Jose, California. 


Available for immediate delivery 


of the largest convention ever held 

Exclusively Devoted to Home & Hobby Computing 
over 300 pages of conference papers, including: 

(Topic headings with approximate count of 7"xl0" pages) 

Friday & Saturday Banquet Speeches (16) 

Tutorials f<>r -the -Computer Novice- (-1-4-) 

People & Computers (13) 

Human Aspects of System Design (9) 

Computers for Physically Disabled (7) 

Legal Aspects of Personal Computing (6) 

Heretical Proposals (11) 

Computer Art Systems (2) 

Music & Computers (43) 

Electronic Mail (8) 

Computer Networking for Everyone (14) 

Personal Computers for Education (38) 

Residential Energy & Computers (2) 

Systems for Very Small Businesses (5) 

Entrepreneurs (6) 
Speech -Recognition & 

Speech Synthesis by Computer (14) 
Tutorials on Software Systems Design (11) 
Implementation of 

Software Systems and Modules (10) 
High-Level Languages for Home Computers (15) 
Multi- Tasking on Home Computers (10) 
Homebrew Hardware (8) 
Bus & Interface Standards (17) 
Microprogrammable Microprocessors 

for Hobbyists (18) 
Amateur Radio & Computers (11) 
Commercial Hardware (8) 


Names & addresses of the 170+ exhibitors at the Computer Faire 

Order mow from: 

Computer Faire 

Box 1579 

Palo Alto C A 94302 


Shipping & Handling: 
Outside California: 

$1 2.00 



Californians Add: 
foside California: 


($1 1.95, plus a nickel, if you prefer) 

(Write for shipping charges ouside U.S.A.) 

Payment must accompany the order. 


6% Sales Tax 

Payment must accompany the order. 

An 8%" x 11" Softbound Book 

O77-1-10 by Computer Faire BOX 1579, PALO ALTO CA 94302 oOo (415) 851-7664 


Jim C. Warren, Jr., Editor 

held in 
The San Jose Convention Center 
San Jose, California 

March 3-5, 1978 


Box 1579 
Palo Alto CA 94302 


^) Computer Faire, Inc. 1978 

all rights reserved 
printed in the U.S.A. 

ISBN 0-930418-01-X 

Library of Congress Catalog card # 78-53026 

These Proceedings were made available, on-site, at the Second West Coast Computer Faire 
because of the heroic efforts of Marc Kindree, Nancy Hamilton, Dave Brown, Gary Markesen, 
and the other humble super-humans working at Nowels Publications, Menlo Park, California; 
because of the super-human efforts of Mort Levine, Gil Anderson, Shirley Boggs, Chris Yanke, 
Ivftte Dawson, John Scroggs, and the other humble heroes working at Suburban Newspaper 
Publications, Cupertino, California; because of the humbling efforts of Toby Forshee of 
Redwood Trade Bindery, Redwood City, California; and, of course, Bill Baumann, Finally, 
the Proceedings could have seen the light of night without the aid of Deft Malloy, and, 
in fact, often did. 


As a widespread movement, "personal computing" began around January of 1975. It 
began as a hobby activity, involving only the dedicated computer hacker and elektroniker 
who had the time, talent, and patience to deal with the relatively sophisticated elec- 
tronics that was available only in kit form, with ~ at most -- minimal documentation, 
and virtually no software. 

Within less than three years, we saw the entry into the marketplace of several fully 
assembled, ready-to-use microcomputers, priced as consumer products for the interested 
technocrat. In noticeably less time than that, we saw the availability of a variety of 
usable — though certainly limited-capability — systems software. 

That is, by 1977, personal computing had moved beyond the dedicated computer hobbyist 
and was beginning to be accessible to the intelligent, logically-oriented novice. 

Now — March, 1978 — we are seeing the first signs of true "computer power for the 
people", as I believe these Conference Proceedings of the Second West Coast Computer Faire 
illustrate . 

In the First West Coast Computer Faire, that took place in April, 1977, we had 
slightly over a day of Conference activities regarding very-low-cost computers in educa- 
tion. This Second Faire has over two days of Conference sessions devoted to the topic. 

Last year, we had two talks concerning the topic that is perhaps the ultimately 
"personal" application of computers — computers for the physically disabled. This year, 
we have a full day of sessions addressing this topic, including demonstrations of several 
operational devices. Additionally, the commercial exhibits include several such demon- 
strations of prototype aids for the physically handicapped. 

In the 1977 Faire, a Conference section addressed the potential of networking per- 
sonal computers. This 1978 Faire — less than a year later — includes details of the 
protocols, and demonstrations of a functioning personal computing network facility. 

Last year, there were few talks concerning the entrepreneur wishing to explore this 
new marketplace, and only one talk addressing microcomputing applications in business. 
This year, half-day sessions address each of these topics, presenting both ideas and the 
results of experience in these areas. 

Though there have been something in the order of 30 other conventions addressing 
the topic of home and hobby computing, to date, the Computer Faire remains unique in the 
fact that it publishes the abstracts and full-text papers of most of the Faire speakers. 
We set this as a major commitment when we created the first Faire; we are continuing that 
commitment for the second Faire. These Proceedings are the result. 

The papers herein were — at most — minimally refereed. As was true of the papers 
in the first Proceedings, they exhibit a wide range in quality. However, they also 
exhibit a timeliness that we feel is essential in a technical area moving as rapidly as 
personal computing is — a timeliness that looks askance at the year-and-more turn-around 
time for obtaining publication in the many heavily-refereed, academically acceptable 
publications. Additionally, these Proceedings illustrate the viewpoint that one need not 
be "academically acceptable" to do interesting and challenging experimentation. They 
also illustrate the view that "novice" is a relative term, and that "state of the art 

I has many dimensions. 

I Jim C. Warren, Jr. 

I Woodside, California 

I 78 February 18 

JIM WARREN, Faire Chairperson 

345 Swett Road 

Woodside, California 94062 
Editor, Dr. Dobb's Journal of Computer 

Calisthenics & Orthodontia 

People * s Computer Company 

Box E 

Menlo Park, California 94025 

ROBERT REILING, Faire Operations Coordinator 


Editor, Homebrew Computer Club Newsletter 

Homebrew Computer Club 

Box 626 

Mountain View, California 94042 


Box 933 

Menlo Park CA 94025 
willing co-pilot for flights of fancy 

1055 Pine 3, Sweet 1 

Menlo Park CA 94025 


Preface, Jim C. Warren, Jr ~ 

Computer Faire Organizers ^ 

Table of Contents -* 


Don't Settle for Anything Less (biographical sketch), Alan Kay 9 

Significant Personal Computing Events for 1978, Adam Osborne ]0 

Dinky Computers Are Changing Our Lives, Portia Isaacson '•> 


Beginner's Guide To Computer Jargon, John T. Shen .... • • • • • • • • • • • • • • • • • • • ;; • • • ; • ; ; * 

Everything You Never Wanted To Ask About Computers Because You Didn't Think You d Understand It Anyway, Ur 

A Talk For People Who Got Talked Into Coming Here By Someone Else, Jo Murray lj 

Introduction to Personal Computing, A Beginners Approach, Robert Moody 24 


Electronics for the Handicapped (brief abstract), Robert Suding * ' 

Microcomputer Communication for the Handicapped, Tim Scully • . . . • • • • : - J* 

Speech Recognition as an Aid To The Handicapped (brief abstract), Horace Enea and John Reykjalin ^ 



Microprocessors in Aids For The Blind, Roberts. Jaqujssjr. . . ^-^ ^__„ _ A AIU _ D A|den 47 


Microcomputer-Based Sensory Aids For The Handicapped, J.S.Brugler 70 

Blind Mobility Studies With A Microcomputer, Carter C. Collins, William R. O'Connor and Albert B. 

The Design of A Voice Output Adapter For Computer, William F. Jolitz . . . . • £° 

Development of Prototype Equipment To Enable The Blind To Be Telephone Operators, Susan Halle Phillips w 

The Design of A Voice Output AdapterFor Computer,JVilliam F. Jolitz 
Development of Prototype Eqi ' "" n " JT " T "' " 

Microcomputer-Based Sensory 


Ambitious Games For Small Computers, Larry Tesler. /J 

Epic Computer Games: Some Speculations, Dennis R. Allison and Lee Hoevel »• V J i/ 1/ u 7fi 

Create Your Own (Computer) Game, An Experience in Synectic Synergistic Serendipity (abstract), Ted M. Kahn /» 

Psychological Tests With Video Games, Sam Hersh and Al Ahumada /y 


Computer Art and Art Related Applications in Computer Graphics: A Historical Perspective and Projected Possibilities, 

Beverly J. Jones 81 

Microprocessor Controlled Synthesizer, Caesar Castro and Allen Heaberlin 85 

Designing Your Own Real-Time Tools, A Microprocessor-Based Stereo Audio Spectrum Analyzer for Recording Studios, 
Electronic Music, And Speech Recognition, Byron D. Wagner 96 


Personal Computing and the Patent System, David B. Harrison 105 

Copyright and Software: Some Philosophical and Practical Considerations, Kenneth S. Widelitz 115 


Becoming A Successful Writer About Computers, Ted Lewis 117 

Writing A User's Guide, Douglas j. Mecham 119 

Editing and Publishing A Club Newsletter, Richard J. Nelson 125 


Deus Ex Machina, or, The True Computerist, Tom Pittman 132 

Peoples' Capitalism: The Economics of the Robot Revolution, James S. Albus 135 

Thoughts on the Prospects for Automated Intelligence, Dennis Reinhardt 140 

Brain Modeling and Robot Control Systems, James S. Albus 144 


A Peek Behind the PCNET Design, Mike Wilber 153 

Communication Protocols for a Personal Computer Network, Ron Crane 156 

PCNET Protocol Tutorial, Robert Elton Maas 159 


Micro's In The Museum: A Realizable Fantasy, Disneyland On Your Doorstep?, Jim Dunion 169 

The Marin Computer Center: A New Age Learning Environment, David and Annie Fox 173 


Personal Computers and Learning Environments: How They Will Interact, Ludwig Braun 177 

Personal Computers and Science Museums(brief abstract), Arthur Luehrman 178 

Computers for Elementary School Children (brief abstract), Bob Albrecht 179 

Bringing Computer Awareness To The Classroom, Liza Loop 180 

Implications of Personal Computing For College Learning Activities, Karl L. Zinn 182 

Getting It Right: New Roles For Computers In Education, Thomas A. Dwyer 193 

The Role of the Microcomputer in a Public School District, Peter S. Grimes 195 


Microcomputers in a High School: Expanding Our Audience, William J. Wagner 198 

Introducing the Computer to the Schoolroom, Don Black 203 

Education or Recreation: Drawing the Line, William P. Fornaciari, Jr 206 

Learning With Microcomputers, Richard Harms 21 1 

Back to BASIC (Basics), David M. Stone 213 

A Comprehensive Computer Science Program for the Secondary School Utilizing Personal Computing Systems, Melvin L. 

Zeddies 216 

Microprocessor Computer System Uses in Education(Or, You Can Do It If You Try), Robert S. Jaquiss, Sr 223 

The Computer in the Schoolroom, Don Black 232 


So You Want To Program For Small Business, Michael R. Levy 239 

Budgeting for Maintenance: The Hidden Iceberg, Wm. J. Schenker 245 

Microcomputer Applications in Business: Possibilities and Limitations, Gene Murrow 254 

MICROLEDGER: Computerized Accounting for the Beginner, Thomas P. Bun 261 


Money For Your Business— Where to Find It, How to Get It, Don Dible 267 

Selling Your Hardware Ideas: How To Start and Run A Manufacturing Oriented Computer Company, Thomas S. Rose 271 
Bringing Your Computer Business On-Line, Stephen Murtha, Elliott MacLennan and Robert Jones 276 


Toward a Computerized Shorthand System, W.D. Maurer 278 

Microcomputer Applications in Court Reporting, Douglas W. DuBrul 285 

Real Time Handwritten Signature Recognition, Kuno Zimmermann 291 

Input Hardware Design for Consumer Attitude Research With a Microcomputer, H.P. Munro 295 

Improving Name Recognition and Coordination in Video Conferencing, David Stodolsky 301 

The Bedside Microcomputer in the Intensive Care Nursery, Robert C.A. Goff 303 

An Automated Conference Mediator, David Stodolsky 307 


Synthetic Speech from English Text (brief abstract), D.Lloyd Rice 317 

Machine Recognition of Speech, M.H.Hitchcock 318 


SSTV Generation by Microprocessors, Clayton W. Abrams 321 

A Real Time Tracking System for Amateur Radio Satellite Communication Antennas, John L. DuBois 325 


Microprocessor Standards: The Software Issues, Tom Pittman 343 

Proposed IEEE Standard for the S-100 Bus, George Morrow and Howard Fullmer 345 


Two Cheap Video Secrets, Don Lancaster 362 

A Recipe for Homebrew ECL, Chuck Hastings 370 

N— Channel PACE 16-bit Microprocessor System, Ed Schoell 383 


Microprocessor Interfacing Techniques, Rodnay Zaksand Austin Lesea 387 

Testing for Overheating in Personal Computers, Peter S. Merrill 390 


Interfacing a 16 Bit Processor to the S-100 Bus, John Walker 394 

Single Chip Microcomputers for the Hobbyist, John Beaston 402 

The Disystem: A Multiprocessor Development System with Integrated Disc-Oriented Interconnections, Claude Burdet. 406 
A Point-Of-Sales Network, Samuel A. Holland 423 


A Short Note on High Level Languages and Microprocessors, Sassan Hazeghi and Lichen Wang 429 

Compiler Construction for Small Computers, R. Broucke 441 

Table Driven Software: An Example, Val Skalabrin 445 

Design Considerations in the Implementation of a Higher-Level Language, William F. Wilkinson 451 

An Arithmetic Evaluator for the SAM-76 Language, Karl Nicholas 460 


ALGOL-M: An Implementation of a High-Level Block Structured Language for a Microprocessor-Based Computer 

System, Mark S. Moranville 469 

SPL/M - A Cassette-Based Compiler, Thomas W. Crosley 477 

An Experimental PASCAL-like Language for Microprocessors, H. Mare Lewis 489 

An Introduction to Programming in PASCAL, Chip Weems 494 



User documentation, internal specifications, 
annotated source code. In the two years of 
publication, DDJ has carried a large variety of 
interpreters, editors, debuggers, monitors, 
graphics games software, floating point 
routines and software design articles. 




Dr. Dobb's Journal publishes independent 
evaluations-good or bad- of products being 
marketed to hobbyists. It is a subscriber- 
supported journal. Dr. Dobb's carries no paid 
advertising; it is responsible only to its 
readers. It regularly publishes joyful praise 
and raging complaints about vendor's 
products and services. 

po box 6528 denvcf, Colorado 80206 (303) 777-7133 


It is not very often that there is a journal/newsletter that the Digital Group 
is able to recommend without some hesitation (and we get them all) . However, 
Dr. Dobb's Journal of Computer Calisthenics S Orthodontia is one pleasant 
exception. Jim Warren, the editor, has put together a good concept and is 
managing to follow through very well indeed. There is no advertising in the 
Journal . It is supported solely on subscriptions. That also means that 
manufacturers have zero leverage over the content of the magazine. The Joqrnal ' s, 
primary purpose is" to-pl ace significant software into the public domain *r.4 to 
provide a communications medium for interested hobbyists. The approach is 
professional and they are growing quickly. 

(In case it might appear otherwise to some people, there is no official link 
whatsoever between the Digital Group and Dr. Dobb's Journal - we've taken our 
lumps as appropriate just like everyone else when Jim felt they werf- justified.) 

We think Dr. Dobb's Journal is here to stay and 
for everyone in the hobbyist world of computers. 

publication that is a must 
Don't miss it! 

THE software source for microcom- 
puters. Highly recommended." 

Philadelphia Area Computer Soc. 
The Data Bus. 

"It looks as if it's going to be THE 
forum of public domain hobbyist 
software development. 

Rating- ft ft & ft" 
Toronto Region Association of 
Computer Enthusiasts (TRACE), 

"The best source for Tiny BASIC and 
other good things. Should be on your 

The Computer Hobbyist, 
North Texas (Dallas) Newsletter 


♦ Hot News& 

Raging Rumor 

* Systems Projects 

Dr. Dobb's journal 

of Computer Calisthenics / & Orthodontia 

Please start my one-year subscription (ten issues) to Dr. Dobb's Journal 
of Computer Calisthenics & Orthodontia and bill me for just $12. 





Unconditional Guarantee: If you ever wish to discontinue your 
subscription for any reason, we will refund the complete amount for 
the remainder of your subscription. 

□ Visa 

□ Mastercharge 

Card Number 

Expiration Date 

Outside the U.S., add $4 for surface postage. Airmail rates on request. 

Mail this coupon or a facsimile to: Dr. Dobb's Journal , Dept 56, 
El Camino Real, Box E, Menlo Park CA 94025 


Alan Kay 

Xerox Palo Alto Research Center 

3333 Coyote Hill Road 

Palo Alto CA 94304 


As a child, Alan Kay found himself equally attracted to the arts and sciences. In 
fact, he has never been able to discover any important distinction between the two. 
A short stint as an illustrator and professional musician was followed by the pursuit 
of mathematics and biology, occasionally interrupted by involvement in theatrical 

Eventually he discovered that the world of computers provided a satisfying 
environment for his blend of interests. A PhD (with distinction) from the 
University of Utah led to a research position at Stanford University and then to the 
Xerox Palo Alto Research Center where he is a Principal Scientist and Head of the 
Learning Research Group. 

In 1967-69, while at the University of Utah with Ed Cheadle of Memcor Inc., he 
designed the FLEX Machine, the first higher-level personal computer. At Xerox he 
started the Learning Research Group, a ten -year project to produce Dynabook, the 
personal computer of the 1980's. He is the initial designer of Smalltalk, the 
programming system of the Dynabook. 

Whenever he can he designs musical instruments, cooks, and plays tennis. 

Selected Writings 

E.LE.K-M o.c.hin?. 

FLEX, A FLexible Extensible Language, Tech. Rep. 4-7, C.S. Dept. U. Utah, 1968 
The Reactive Engine, PhD Thesis, C.S. Dept U. Utah, 1969 

Early Dy naboo k and Smalltalk 

A Personal Computer for Children of All Ages, ACM Nat'I Con., Boston, Aug 1972 
A Dynamic Medium for Creative Thought, NCTE Nat'I Con., Minneapolis, Nov 1972 

Vintage Dy naboo k and Smalltalk 

Personal Computing, Con. 20 yrs of Com. Sci., U. Pisa, Italy, June 1975 
Personal Dynamic Media, w/ A. Goldberg, Xerox PARC (1975) 

" , w/ A. Goldberg, exerpts: IEEE Computer, Mar 1977 
Teaching Smalltalk, w/ A. Goldberg, Xerox PARC, June 1977 
Microelectronics and Personal Computers, Scientific American, Sept. 1977 



Adam Osborne, President 
OSBORNE & ASSOCIATES, INC., 6 30 Bancroft Way, Berkeley, 

CA 94710 


This paper examines the princi- 
pal microprocessor achievements of 
1977, and forecasts significant events 
for 19 78. The emphasis is on semicon- 
ductor parts that have been developed 
rather than on home computing system 
hardware or software. The three most 
significant parts to be developed and 
shipped in 1978 are identified. 

The recipient of the White Ele- 
phant Award for achievement and per- 
sonal computing will be announced at 
the dinner. This award is described 
in the paper. In order to be consis- 
tent with the strange logic of the 
semiconductor industry, the White 
Elephant Award is an award for out- 
standing achievement rather than an 
award for lack of achievement, as the 
name might suggest. 

Significant Developments from 19 77 

I would like to summarize what 
I believe to be the most significant 
microcomputer industry achievements 
of 1977, while looking at implica- 
tions for 1978. 

At the level of semiconductor 
components, 1977 was a remarkable 
year in terms of product announce- 
ments and a pretty good year in 
terms of products actually being 
shipped. Let us look at the sig- 
nificant semiconductor developments 
of 1977. 

In 1977 the one-chip, 8-bit 
microcomputer became a reality. 
Mostek started to ship the 38 70 - 
a one-chip F8 - in volume. Intel 
followed closely behind with the 
8048 family of one-chip microcom- 
puters. The 804 8 family is remark- 
able for the presence of the 8 74 8 
series, which provides erasable 
programmable read-only memory on 
the microcomputer chip. This is 
a very significant industry first. 

The 8041 and 8741 are varia- 
tions of the 8048 that need to be 
specially identified. A casual 
reading of data sheets might lead 

one to believe that the 8041 and 8741 
are simply ; variations of the 8048, 
aimed at some obscure corner of the 
market. Nothing could be further 
from the truth. The 8041 and 8 741 
are significant devices because they 
have clearly filled a need. Let us 
explore this need. The concept of the 
one-chip microcomputer was easy enough 
to grasp. Based on the high sales of 
the two-chip F8 configurations, the 
economics of having a very low-cost, 
high-volume, low part-count microcompu- 
ter were self-evident. But this one- 
chip microcomputer provides a small, 
isolated logic system that may well 
exist on its own. A more subtle and 
troublesome problem is the sub-logic 
function, characterized by the device 
controller. It is easy enough to iden- 
tify device controllers such as floppy 
disk controllers, etc. Any microcompu- 
ter system will contain one or more 
of these peripheral devices, each of 
which needs its own interface logic. 
Unfortunately, this interface logic 
must usually be custom designed, re- 
sulting in support functions costing 
far more than the Central Processing 
Unit. This is a problem which is more 
significant than might at fir s-t- appear , 
since microprocessors are being used 
in such a wide and varied set of cir- 
cumstances. Thus, we are not simply 
talking about peripheral devices such 
as floppy disk printers and video dis- 
plays - we are talking about an endless 
and probably unknown set of interfaces. 
The 8041 and 8741 address themselves 
to this sub-logic, interface market. 
Irrespective of what the CPU and the 
peripheral may be, an 8041 will gene- 
rate the necessary interface "intelli- 
gence, providing this interface intell- 
igence can work within the speed, memory 
and I/O constraints of the 8041. To 
complete the effectiveness of the 8041, 
the 8741 allows you to generate inter- 
faces (initially in low volume) by using 
an erasable programmable read-only 
memory to hold programs as they are 

We select the 8741 as the most 
significant part to be introduced and 
shipped in 1977. 



BOX 1579, PALO ALTO CA 94302 

The next area of significant 
development has been the 16-bit 
microprocessor. Fairchild introduced 
the 9440 and started to ship this 
microprocessor, while Data General 
introduced and started to ship the 
MicroNova. Both the MicroNova and 
the 9440 are one-chip implementa- 
tions of Data General Nova Central 
Processing Units. The MicroNova is 
an implementation of the Nova 3/ while 
the' 9440 is an implementation of the 
Nova 1200. 

Specialized processors have 
also begun to appear. Advanced Micro 
Devices has introduced the Am9511, 
which is an arithmetic processor. 
This very significant device finally 
makes it practical to use micropro- 
cessors in intensive computation 
applications. The Am9511 brings 
trigonometric functions, logorithms, 
exponentials and multiprecision 
arithmetic to microcomputer systems. 
We select the Am9511 as the second 
most significant part to be introduced 
during 1977. 

The next area where we have seen 
very significant developments is in 
support circuits for microprocessors. 
A wealth of parallel I/O devices, 
serial I/O devices, DMA controllers, 
priority interrupt controllers and 
peripheral interface circuits were 
introduced. We believe the most 
significant interface circuit to be 
introduced and shipped is the Z80 SIO 
device. The Z80 DMA device should 
also be mentioned, but Zilog is not 
yet shipping it. 

In 1977, Mostek became the first 
company to start shipping 16K-bit dy- 
namic RAMs in volume. Here again is 
a development whose significance can 
easily be overlooked. Why get excited 
about just another memory device? 
Very large, low-cost memory devices, 
as they appear in the future, are like- 
ly to revolutionize more industries 
than any other single development. 
I single out the music industry - the 
recording and reproduction of sound - 
as the one likely to experience deva- 
stating changes in the future. 

Although 1977 was a year for 
announcements and product releases, 
1978 is likely to see even more dra- 
matic new microprocessor-related pro- 
ducts. Specifically, 1978 will be the 
year of the 16-bit microprocessor - 
with the announcement and delivery of 
Intel 8086's, Zilog Z8000's, and sub- 


stantial deliveries of TMS9900's and 
Fairchild 9440* s. Given these develop- 
ments, what impact, if any, can we 
expect on personal computing? 

The answer, surprisingly, is very 
little. Even now, three years after the 
first home computers appeared, there is 
a crippling shortage of software, even 
to support 8080-based microcomputers. 
If a manufacturer were to switch in 1978 
to a new 16-bit microprocessor, it is 
likely to be three or four years before 
this new microcomputer system has any 
reasonable amount of software support. 
Thus, the "software prop" is likely to 
keep existing microcomputers in. commer- 
cial .production for many, many years to 
come. This "software prop" will be re- 
enforced by the fact that , for many app- 
lications, the existing 8080-based micro- 
computer systems are more than adequate 
in terms of computing power; any switch 
to more powerful microcomputers would 
have little tangible economic advantage. 
Even for those applications where more 
computing power is needed, there is 
always the alternative of moving to new 
8080A Central Processing Units that are 
faster - and therefore more powerful - 
rather than moving to entirely new micro- 
processors and instruction sets . 

It is easy to fall into the trap 
of looking upon new microprocessor 
products as "new waves" which replace 
everything that came before them. I 
believe this is a very inaccurate 
visualization of reality. It is more 
accurate to think of new microprocessor 
products opening up new markets - for 
which older microprocessor products 
were inadequate. Once some particular 
level of microprocessor product has been 
adopted, it will be used for a long time 
to come because the cost of re-engineer- 
ing to take advantage of new, more recent 
developments is simply not realistic. 
That is to say, new personal computers 
were manufactured when 8080A Central 
Processing Units and support circuits 
made them economical in the first place. 
Since 8080A Central Processing Units 
and support circuits were adopted in 
personal computers, they will be the 
mainstay of personal computing for many 
years to come. The fact that an 8086 
will be available in 1978 does not mean 
that three years from now all 8080A- 
based systems will be obsolete. Far 
from it. The 8086 is going to have to 
make its own new markets, and will have 
little impact on established markets 
for past microprocessors. Therefore, if 
you are looking at the personal computing 
industry and deciding when to jump in, 

11 BOX 1 579, PALO ALTO CA 94302 

your answer is: as soon as you find 
products you can use. Do not wait 
until next year for better products 
which may appear, because next year 
you will be waiting for the follow- 
ing year, and you may finish up wait- 
ing forever. 

The fact that new developments 
will not cause old developments to 
become obsolete is made more certain 
by the huge customer base for per- 
sonal computing products which 
already exist. The personal compu- 
ting market buoyancy is attested to 
by the present show, and by the 
success of so many other shows 
around the country. This success 
has resulted from a combination 
of eager customers and willing 
visionaries who had the foresight 
to see what was coming and the 
vigor to help it on its way. My 
principal purpose tonight is to 
recognize the individual who I 
believe has done more in the past 
year to further personal computing 
than anyone else. The name of 
this individual will be announced 
at the dinner and not in this 
paper . To this individual , I 
plan to present a singularly apt 
award. In order to be apt, this 
award must recognize the perver- 
sities of the semiconductor in- 
dustry. Instead of rampant infla- 
tion, this is an industry of ram- 
pant deflation. Instead of pro- 
tecting every new product from 
competition, this industry "runs 
out to find a second source, who 
is given all necessary secrets to 
compete effectively. Since every- 
thing is back-to-front in this 
industry, it is only appropriate 
that an award for achievement be 
given a name more aptly associated 
with lack of achievement. There- 
fore, the annual award which I plan 
to present will be known as the 
White Elephant Award. But, instead 
of representing the biggest waste 
of effort, my White Elephant award 
will recognize the best-spent effort. 
The award consists of an 8741 chip, 
which is my choice for Chip-of-the- 
Year, mounted on a suitable plaque 
with a microscopic White Elephant 
cemented onto the surface of the 
chip. I plan to award this trophy 
annually, using the Chip-of-the- 
Year for each year's trophy. I 
furthermore plan to choose the chip 
and the recipient of the award 
entirely on my own, without letting 
my judgment be clouded by committees 


or input from the personal computing 

In recognition of*" the individuals 
who made the selected chip possible, 
the award will list these individuals 
in addition to the person receiving 
the award. 


BOX 1579. PALO ALTO CA 94302 


Portia Isaacson, The Micro Store 

634 S. Central Expressway, Richardson, TX 75080 

(214) 231-1096 

Computers can now (or will soon be) 
found in cars, sewing machines, tombstones, 
typewriters, and pinball machines. The age of 
the abundant computer is here. As it com- 
pletely unfolds we will think we have entered 
a land of science fiction. Dinky computers 
will permeate virtually all aspects of our 
lives. Computers will be used in old ways by 
people and businesses who couldn't afford them 
before and in many exciting new innovative 
ways that we couldn't even have thought of be- 

Computers have been around for some time. 
Why all the fuss now about change? The answer 
is simple. We now realize that computers can 
be useful to individual people. A few years 
ago the price of a computer dropped past a 
threshold that caused a lot of people to under- 
stand that the computer was a personally use- 
ful tool. A few people understood before, but 
now that idea is so popular that it has some 
of the aspects of a religion. The idea of the 
personal computer certainly has a large and 
active following. 

The changes brought about by dinky com- 
puters will be many and not all will be good. 
Change will be rampant in the computer in- 
dustry. But few institutions or individuals 
will escape without change. Businesses both 
large and small, the U.S. economy, labor, 
women, the handicapped, the data processing 
professional, government, the U.S. Postal 
Service, and our educational system are among 
those that will be changed by dinky computers. 

B usiness, Labor, and the Economy 

Small businesses can make use of dinky 
computers in a variety of ways — most of them 
scaled down versions of the same applications 
in big businesses. Applications common to 
most small business include: general ledger, 
accounts payable, accounts receivable, pay- 
roll, and inventory control. Some businesses 
will find a use for word-processing in the 
generation of letters and reports. Mailing 
list maintenance and label generation are 
popular computer uses. A small business 

Of the many words Ted Nelson has given us, 

this is one of the best. 



might find a computer useful in scheduling 
people or equipment. Some businesses will 
have applications specialized to their own 
business such as a personnel agency's main- 
tenance and search of an applicant data base 
or a savings and loan company's calculation of 
amortization schedules. Innovative appli- 
cations might include sales forecasting, elec- 
tronic mail for ordering, building security, 
energy conservation, games as sales techniques, 
and graphics in advertising displays. 

A typical configuration for a small 
business computer system including 32K bytes 
of memory, dual floppy disks and a continuous 
forms printer costs less than $5 per day when 
amortized over three years. Small businesses 
commonly find that a computer costing less 
than $5 per day can replace one or more 
employees and can give the management more 
timely and accurate information than they 
were getting before. In general, the effect 
of the computer on the small business is to 
improve productivity while reducing costs 
primarily by reducing the number of employees 
in relatively unskilled positions. By re- 
ducing overhead an increasing number of small 
businesses will find themselves viable. This 
experience is not unique to small businesses 
but is the same as that of large corporations 
which preceded them in the use of business 
computers. Future applications could include 
conferencing and working at home. 

The effects of the managers' use of the 
dinky computer will be many. The productivity 
of clerical employees will be increased. The 
effect of an easily accessible private com- 
puter will be to improve budgeting and project 
control techniques. Electronic mail will de- 
crease the need for unskilled labor and de- 
crease the use of the post office. 

The same $5 per day business computer 
system found so helpful in small businesses 
will also be useful to the manager in the 
large corporation. Now a manager at nearly 
any level can afford his or her own private 
computing resource. One of the first appli- 
cations will be word processing for the 
preparation of letters, memos, and reports. 
Other immediate applications include: 
budgeting, project control, maintenance of 
specialized data bases, sales forecasting, 

BOX 1579, PALO ALTO CA 94302 


scheduling* reminders, mailing or routing list 
maintenance, and electronic mail. 

The overall effect of the use of low-cost 
computing in business will be an increase in 
.national productivity and an improved economic 
position for the U.S. in the world marketplace. 
The U.S., as the undisputed leader in low-cost 
computing technology, will be able to use this 
technology as a principal weapon in any future 
economic war. 

The labor force will experience both 
positive and negative effects of low-cost com- 
puting. On the positive side, there will be 
reduced need for people to do boring work. 
However, there will be a reduction in the de- 
mand for relatively unskilled labor such as 
clerical, mail service, and bookkeeping. 
Since most of these jobs are now filled by 
women, women will be hardest hit by the re- 
duced demand for unskilled labor. Countering 
the increasing demand for programmers will be 
the fact that entry-level programmers will be 
in plentiful supply since low-cost computing 
will make computer education, even self- 
education, widely available. 

Now computer-inventiveness is in the 
public domain. Before only large corporations 
and well-endowed universities could invent 
products containing computers. Now the man or 
woman on the street has economic access to 
computers and can use them in inventions. 
I'm sure they will. The same inventive talent 
that brought us the automobile and the elec- 
tric light will bring us "intelligent" compu- 
ter-based products that are now beyond our 
imagination. The businesses springing up 
around these inventions will employ people and 
further Improve the U.S. economic position. 

The Computer Industry 

As the demand for dinky computers goes 
up, the demand for gargantuan computers will 
come down. It will often be found that new 
applications, or portions of new applications, 
are more economical on small computers. The 
traditional corporate demand for bigger and 
bigger computers will slacken as fewer new 
applications are developed for it. Addi- 
tionally time-sharing use of the big corporate 
computer will be replaced by small computers 
in instances that are not locked in by data 
bases or applications software. 

The corporate data processing center 
will lose control of the data processing 
function as more and more departments own 
their own computers. The DP center will do 
less new development since new projects will 
be done at the department level if possible. 
The DP center may find a new role when depart- 
ments realize that they want to access the 
central data base and communicate with the 
computers of other departments. DP's new 
role will be in planning the distributed 



data base and communications networks. This 
role will not be easy since departments will 
realize that information is power. The 
struggle over how to distribute the data base 
will be a power struggle between departments 
with DP caught in the middle. 

Now that computers can be owned by indi- 
viduals or dedicated to the use of an indi- 
vidual in a corporation, there is little need 
for time-sharing. In fact, time-sharing was 
invented as an attempt to give the illusion 
that each user had his or her own computer. 
Now that each user can have his or her own 
computer, time-sharing is no longer needed and 
the overhead required by sharing makes it un- 
competitive. Present time-sharing customers 
will, of course, stay with time-sharing if 
they are locked in by software or data bases. 
Additionally, there are a few applications 
that may need resources too great for today's 
dinky computer. 

The big computer will not go down with- 
out a fight. We can expect to see signifi- 
cant price cuts in order to keep the 
gargantuan machine alive. But ultimately 
the giants will be kept only to run programs 
too hard to change. Most new architectures 
will be based on unshared computers, shared 
large disks, and shared fast peripherals 
connected into networks. The heyday of dis- 
tributed computing will have arrived. 

The new computer industry will see many 
opportunities. Computer manufacturing and 
distribution will be feasible small businesses. 
The new small companies with low overhead will 
keep the price of computing low; and, in fact, 
may provide the solution to the problem of the 
present near -monopoly in the industry. There 
will be a new economics associated with mass 
produced software. A complex software package 
may sell for just a few dollars because it will 
be sold thousands of times. Individuals may 
be able to capitalize on their efforts in 
software creation' through royalty payments in 
much the same way as authors of books do now. 
The data processing professional will be 
faced with many changes. The data processing 
department will need maintenance programmers, 
communications and network experts, and data 
base designers. Programming will be done in 
user departments where application knowledge 
will be at a premium. So programmers who 
don't fit into the new DP department will find 
themselves in user departments specializing 
in a particular application area. This 
specialization will certainly limit their 

Although lower-cost computers will mean 
more computers and a great demand for pro- 
grammers, the greater demand will be offset 
by a greatly increased supply of entry-level 
programmers and the fact that programming 
will be easier. Schools at all levels will 
be able to offer computer training since the 

BOX 1579, PALO ALTO CA 94302 

hardware is now affordable. Many people will 
even teach themselves how to program. The 
new dinky computers are interactive and much 
easier to program than big batch computers. 
All this could lead to a decrease in the 
salary-level of entry-level programmers. Ul- 
timately this must affect other levels. 

As the public becomes more and more 
knowledgeable about computers, the job of the 
data processing professional will seem much 
less glamorous and mysterious and much more 
just an ordinary job. This will have more 
than just an ego deflating effect on the pro- 
fession. A computer-literate public will de- 
mand that the programming job be done properly 
with the good of the public an objective. We 
can expect to see a public demand for legis- 
lation to control computer usage and program- 
mer qualifications. As the public becomes 
more aware that they are becoming increasingly 
dependent on unproven computer technology, our 
profession may find itself in the fish bowl 
of public controversy. 


Government at all levels will experience 
most of the problems and opportunities of 
businesses. In addition, government will face 
some unique changes. The increasing use of 
electronic mail will bring about further de- 
clines in the use and efficiency of the U.S. 
Postal Service. Government may be able to re- 
duce the demand for energy by encouraging the 
use of computers to control and conserve ener- 
gy usage in homes and industry. Crime can be 
decreased through the use of computerized 
security systems. The cost of political 
campaigns may be decreased by applying low- 
cost computing to the data processing tasks 
involved in a campaign. Government must help 
solve the problem of protection for the 
author's rights in mass-produced software. 
Increasing displacement of unskilled labor by 
computers will be a difficult governmental 
problem. New legislation may be required 
to control computer technology. Finally, our 
government will be faced with the new ghetto 
of the computer "have-nots." 

The Individual 

All the changes previously mentioned 
affect us to some extent individually. There 
are other effects, however, that deserve 

The computer brings us a new form of 
entertainment. It is entertainment through 
the simulated experience. Often called com- 
puter games, this form of entertainment can 
offer very challenging and highly involving 
activities. The most popular game of this 
class is Star Trek. It lets one pretend to 
be captain of a star ship charged with 

defending the universe against klingons. 
The strategies and events are intricate and 
demanding requiring quick and correct de- 
cisions. Computer games are often intel- 
lectually stimulating as well as just plain 
fun. Although the computer games encourage 
socialization to an even less extent than 
television (there are no commercials), at 
least they involve the player in the activity 
unlike passive television-watching. 

Besides games, the computer offers other 
opportunities for entertainment and creativi- 
ty via computer-generated art and music. For 
several years a few artists and musicians 
have experimented with the computer as a tool 
for creativity and expression. Now the com- 
puter as an artist's tool is available to 

The low-cost computer coupled with video 
disk technology could do much to increase the 
availability and flexibility of personalized 
education. These new technologies make high- 
quality computer-assisted instruction techni- 
ques affordable by educational institutions, 
libraries, corporations, and individuals. 
The place of education may become much more 
flexible. The role of the educational in- 
stitution may change to primarily that of 
preparing courseware and certification of 
knowledge or skill levels. 

Computers can be used in many ways to 
improve the lives of the handicapped. A 
person without arms or legs could control a 
wheelchair by voice commands. .A blind person 
might use a typewriter, computer terminal, 
or calculator that speaks each letter or 
number. A deaf person might use a telephone 
that visually display es messages. A speech- 
impaired person might use a speech synthesis 
device that spoke what was entered at a key- 
board. The possibilities are exciting and 

In the gizmo age we will be surrounded 
by "intelligent" devices ranging from the 
self -dialing phone to the self -flushing toilet. 
Most of these devices will be helpful and 
friendly, but not all. The computer -generated 
junk phone call is with us. A computer-based 
device can place calls, play a recorded 
message, record a response, and even accept 
touch-tone input of a credit card number for 
a purchase. The unlisted number doesn't help 
since the device could place calls to all the 
numbers having a certain prefix — a very in- 
expensive way of placing calls to a part of 
town corresponding to a certain economic 
level. The devilish device could remember 
that you didn't answer and call until you do. 
It could even remember that you hung up and 
pester you until you listen. Unfortunately, 
junk telephone calls are a fraction of the 
cost of junk mail. A bill has already been 
introduced in the Congress to control this 
nuisance made possible by dinky computers. 



BOX 1579, PALO ALTO CA 94302 

What will be next? 

Low-cost computing will add fuel to the 
already threatening invasion of individual 
privacy. Abundant dinky computers mean data 
bases too numerous to control. An indivi- 
dual won't have a chance of knowing whose 
keeping what records about him or her. Cheap 
computers will mean increased feasibility of 
surveillance of individuals by government or 
business. The IRS might be able to check, in 
detail, every tax return. Isn't that exciting! 


We've surely only glimpsed the brave, new 
world being created by dinky computers. The 
next few years will be more exciting and 
probably less believable than most science 
fiction. I want to be there as it happens. 
Perhaps I can help 


John T. Shen 
Computer Scientist & Consultant 
Naval Ocean Systems Center 
271 Catalina Blvd 
San Diego, CA 92152 

Human nature being what it is, we're always trying 
to develop tools to make problem-solving easier. 
We also try to develop tools to do monotonous and 
mechanical jobs so we HI have more free time to do 
the things we enjoy. So, unlike humans, the tools 
we develop make fewer mistakes, work without getting 
tired and don't go on strike. 

One of the best tools we've created is that 
creature called "computer." But what is a computer? 
What's the difference between a computer and a 
microcomputer? What do we mean by multiprocessing? 
And what do we mean by large scale integration 


A computer is an electronic tool that can 
accept Information supplied by a human or another 
machine, A computer also accepts instructions 
regarding what to do with the information 
supplied. The computer then performs the 
operations on the given information. After the 
instructions are performed, the computer 
supplies the results to the person who requested 
them, or to another machine which may need the 
results to carry out other operations. 

A basic computer is usually composed of 
an input and output (I/O) unit, memory Cor 
"storage") unit. 

The input unit accepts the information to 
be operated on from people or other machines, 
and the output unit makes the results available 
in terms a human can understand. 

The memory unit stores information until 
needed by one of the other units, such as the 
arithmetic and logic unit, the control unit 
or the I/O unit. 

The arithmetic and logic unit CALU) does 
the arithmetic and logic operations necessary 
to sort or search for particular items or 
perform mathematical procedures. 

The control unit manages all the other 
units. For example, the control unit decides 
when the I/O unit will accept information and 
when the information should be sent from the 
I/O unit or the memory unit to the ALU for 
processing. The control unit also decides 
what operations to carry and in what sequence. 
When an operation is completed, the control 

unit decides whether the results should 
be sent to the I/O unit or the memory 

unit. . ., .. ^ j i„ 

The technology for building today s 
computers is very different from the 
technology that built computers 10 to 
20 years ago. Until the late 1950's, 
computers were built from electronic 
tubes, mechanical relays, resistors 
and capacitors. We call these computers 
the "first generation". 

From the late 1950's to the early 
1960's, computers were built from 
discrete transistors, resistors and 
capacitors. We call these computers 
the "second generation"^' 

In the early 1960's, a new tech- 
nology arose, called integrated circuits 
(IC), where many components are fabri- 
cated on a chemical substrate called a 
"chip", which is about 1 centimeter 
square. In the early 1960's only 100 
transistors could be packed on a chip. 
Computers implemented with 100-transis- 
tor chips are called the "third-genera- 
tion". . ,. u • , 
Later, new fabrication techniques 

were developed, so that today we can 
pack 1000 or more transistors on a 
single chip. We call these computers 
"fourth generation". 

From the first to the fourth 
qeneration, the physical size of a 
computer with the same computing power 
shrank drastically. The cost also 
decreased impressively, and the com- 
puting speed increased several magni- 
tudes. _. X 

A computer designed for use in 
many fields of business and science is 
called a "general -purpose computer . 
A computer designed for a specific 
purpose, such as monitoring a patient s 
heart condition, is called a special- 
purpose computer". 

A small computer is called a "mini 
computer". A very small computer is 



BOX 1579, PALO ALTO CA 94302 

a "microcomputer". The central processing unit 
(CPU) of a microcomputer is called a "micro- 

If a computer has more than one CPU, and if 
the CPU's are operating in parallel, the computer 
is called a "multiprocessing computer" or a "multi- 

But just having a computer will not solve 
your problems. You need a way to instruct the 
computer to solve a problem. One way is to write 
a "program" (a set of instructions or steps that 
tell the computer exactly how to solve a problem. 

Since English is our native language the 
languages used for writing programs are usually 
English-based, examples of English-based languages 
are FORTRAN, COBOL and ALGOL. Some of the pro- 
gramming languages are mathematically-oriented, 
such as APL. Both types are "human under- 
standable". They are called "high-order languages". 

But all computers are built on the simple 
language of "yes" and "no" or (Ts and o's), which 
is the "machine language". To use high-order 
languages, we must build translators to act as 
interpreters between man and machine. 

The programs programmers write to solve 
their particular problems are called "application 
programs". The large program developed by the 
computer manufacturer for managing the computer 
resources such as I/O devices, memory spaces and 
CPU time is called the "operating system". 
Programs that facilitate the easy use of I/O 
devices and peripheral memory are called 
"utility programs". 

The application programs, language trans- 
lators, operating system and the utility pro- 
grams are called "software". 

When a language translator completely trans- 
lates a program before the execution of that 
program, the translator is called a "compiler". 
If a translator translates one statement of a 
program at a time and executes that statement 
immediately, the translator is called an 

When a portion of a control unit is 
electrically programmed into a device call 
"read-only memory' (ROM), or when some of the 
software is electrically programmed into one or 
several ROMs, the programming is called "micro- 




Copyright Jo Murray, 1978 

Jo Murray 
2325 Lelmert Blvd. 
Oakland, Ca. 94602 

This talk will trace the history of 
computers, beginning with Charles Bab- 

to ignore them. 

Actually, when computers were first 
invented, Just about everybody did ig- 
nore them. Hard as it is to believe, 
there was a man back in the early nine- 
teenth century who invented all the 
principles of computers. His name was 
Charles Babbage, and he got a lot of 
help in interpreting his ideas from the 
Countess of Lovelace. You may know her 
better as the only daughter of the 
poet Lord Byron. Other people call her 
the first computer programmer. 

Her parents separated Just after she 
was born, and she never saw her father, 

wwim..™-., —c* ~o but * e did wrlte her l Qtter8 ^d he „ 

bage, who might have given us computers apparently referred to her in some of 
more than 100 years ago if he had only his poems, although he didn't seem en- 
known more about electricity. Bab- tirely pleased with her intellectual 
bage was aided by the Countess of Love- talents. A lot of people think he may 
lace, the daughter of that soulful and have had her in mind in the passage in 

anti-machinery poet Lord Byron. Then 
there was George Boole, another nine- 
teenth century figure who gave us the 
symbolic logic that lets computers de- 
cide what to do next. The history 
continues into the current century with 
a few references from such technical 
publications as Alice Through The Look - 
ing Glass , a brief description of 
vacuum tubes, transistors and the sil- 
icon chips that run the computers at 
the Falre, and no formulas whatsoever. 

If you're one of those people who 
have been thinking that alii computers 
do is print out bills in funny look- 
ing numbers and letters and then con- 
fuse your orders with your next door 
neighbor's, then this is the place for 


You know, you Just kept thinking 
that if you ignored them long enough, 
they'd go away. Then the humans who 
could do thinks like talk on the tele- 
phone and read names and addresses in 

Don Juan which reads: 

"'Tis pity learned virgins ever wed 

With persons of no sort of education 

For Gentlemen although wellborn and 

Grow tired of scientific conversation. 

I don't choose to say much upon this 

I am a plain man and in a single 
8 tat ion 

But, Oh ye Lords of ladies intel- 

Inform us truly, have they not hen- 
pecked you all?" 

Other people thought a little more 
of her intellectual talents. When she 
first met Babbage, she was with a group 
that was described as looking at his 
machine as if they were a bunch of 
savages looking at a gun for the first 
time. But Lady Lovelace apparently 
grasped the principles of Babbage* s 
machine the first time she saw it. 

longhand instead of making you put them She even predicted that some day the 

in block letters in little squares 
would come back. 

But so far they haven't. And if 
you're like me, one day you decided 
it is convenient to have computers 
that know whether there are seats on 
planes and telephones you can use to 
call across the country and little 
calculators you can hold in your hand. 
That's not even considering all the 
wonderful machinery here today. 

Not all of the electronic marvels 
I we have today are computers, strictly 
I speaking. But it is getting very hard 

engines would be used to write music. 

Babbage got the idea because he was 
fed up with the mathematical mistakes 
of the times. In his day, sailors and 
astronomers and anyone else interested 
in math carried huge books of tables. 
But they were calculated by hand and 
they were set in type by hand, and the 
number of mistakes was incredible. 
One book of calculations for sailors 
was so bad that the captain of one 
ship, who got it as a gift and didn't 
realize it was more vaulable for its 
beautiful bindJqg than its accuracy, 



BOX 1579, PALO ALTO CA 94302 

was never heard from again. 

So Babbage sat down and Invented 
his difference engine, as he called 
it. It worked by calculating tables, 
using the difference between two num- 
bers. The idea was that if one pound 
of meat cost five shillings, two pounds 
would cost ten shillings and three 
pounds would cost fifteen shillings. 
So instead of multiplying three times 
five to find out how much three 
pounds of meat would cost, you could 
look at a table that was constructed 
by adding five each time. 

The machine was also capable of 
making tables involving squares of 
numbers, using the principle of the 
second derivative, or "difference." 

This principle was not new: this 
was the way most mathematical tables 
were constructed at the time. What 
was new was the idea of having a 
machine do it, and the idea that a 
machine could be constructed so it 
would never make mistakes. 

Babbage used punch cards, which 
had been developed in France to con- 
trol looms so they would weave pat- 
terns in cloth, to feed his machine 
information. He even devised an in- 
genious system, based on logarithms, 
so the machine would stop and ring a 
bell if the attendant gave it the 
wrong card. The engine printed out 
the answers itself to eliminate the 
possibility of mistakes in typeset- 
ting. He then went on to design a 
more sophisticated machine, which he 
called the analytical engine, that 
would do almost everything computers 
do today. 

Babbage* s principles were so close 
to today's that Howard Aiken, who 
helped build one of the early modern 
computers, once said, "If Babbage had 
lived 75 years later, I would have 
been out of a Job." 

Babbage, who held himself in very 
high esteem, apparently agreed. He 
wrote that if anyone later developed a 
similar machine, "I have no fear of 
leaving my reputation in his charge, 
for he alone will be able to fully 
appreciate the nature of my efforts 
and the value of their results." 
When Aiken came across these lines, 
he said he felt like it was a voice 
personally addressing him from the 

But Babbage had two things going 
against him, and his engines were 
never completed. 

For one thing, electricity had 
Just been discovered. It was only 
in 1831 that Michael Faraday discov- 

ered a moving magnet would Induce an 
electrical current in a coil of wire. 

For another thing, toolmaking was 
a relatively new art. Clocks were 
about the most complicated machines 
that existed, and they were all Individ- 
ually made by hand. Since hecould not 
use electricity, his machines needed 
an enormous number of gears. He 
had to have a workman make most of 
the tools he needed to make the pre- 
cise gears, and then his chief workman 
got mad and quit and ran off with the 
tools. But that was another problem. 

Babbage did gain a sort of prominence 
for his time. There were Just enough 
people who appreciated his genius that 
when he died in 1871, the Royal College 
of Surgeons of England preserved his 
brain, which is still there today. 
But the surgeons who examined it to 
see if brain looked different from 
anybody else's couldn't find anything 
especially remarkable about it. 

After Babbage and the examination 
of his brain, computers faded from 
the scene for awhile. There we»e Just 
a number of minor steps that all had 
to be taken before the first modern 
computers could be built during World 
War II. 

For one thing, Babbage wanted his 
engine to be smarter thanmost computers 
are today. He wanted it to be able 
to do sophisticated things like add 
204 and 31 1 . Both of his engines— 
which were never completed--were to 
do this using the decimal system in 
the same sort of way that the odometer 
on a car ¥OPk5. When one row of fig- 
ures reached 10, it was to automatical- 
ly cause the wheel of figures in the 
next row to turn. 

Today's computers don't even try 
to count as high as 10. They're like 
the Red Queen in Alice Through the 
Looking Glass . You remember when she 
asked Alice, "What's one and one and 
one and one and one and one and one 
and one and one and one?" 

"I don't know," Alice replied. "I 
lost count." And then the Red Queen 
yelled, "She can't do addition." 

Well, the difference between us 
and computers is somewhat like the 
difference between Alice ad the Red 
Queen. Just about anybody here can 
add 204 and 3 11 in their heads if 
they put their minds to it. But if I 
stood here and said "one" over and 
over again, first for 311 times and 
then for 204 times, I doubt that any- 
body could tell me exactly how many 
times I said "one." 

Well, this is what computers do. 



BOX 1579, PALO ALTO CA 94302 

They say one, one, one, one, one, one, 
one, etc. and they keep track of it. 
Or they subtract *one* two or three hun-i 
dred or thousand times. And if you 
think about it, multiplication is 
simply a matter of adding numbers and 
division is simply a matter of sub- 
tracting the divisor over and over. 

Actually, it's a little more compli- 
cated than this, but this is basical- 
ly how they work. The don't know any 
numbers but ones and zeros and they 
count them over and over again. 

This is known as the binary system, 
and when you take the binary system 
and electricity, you can do some in- 
credible things with computers. 

The reason we use 10 as a base is 
probably because we have 10 fingers. 
Babbage used base 10, too. But the 
modern computers don't have 10 fingers 
or even 10 rows of digits like the 
early machines. They Just have elec- 
trical switches. They're either on 
or off. They either have current flow- 
ing through them or they don't. 
If they're on, the computer counts 
them as one. If they're off, the com- 
puter counts them as zero. This gives 
you a numbering system that' s very 
easy for computers, even though it's 
difficult for humans. 

Probably people who work with com- 
puters a lot can look at binary fig- 
ures and read them as easily as we can 
read the decimal system. But I can't. 
So I'm going to refer to this little 
card to explain the binary system. 

The digit in the righthand column 
represents the number of ones. The 
other columns are not powers of 10, 
but powers of 2. 








4 . 
1 1 












2 (2]+0) 


is the 3 (2U1) 


same 4 (2 2 +0+0) 


as 5 (2| + 0+l) 


6 (22+2^0) 


7 (21+2U0) 


8 (2 5 +0+0+0) 

You can add them Just like you 
add in the decimal system, except 
that as soon as the total is 2, 
you have to carry a digit to the 
next column. 

You can also use the on-off switches, 
or the zeros and ones, to represent 
letters. You can say A=0, B=1 , C=01 , 
and so on. By the time you get up to 
five digits, you have 32 different 
combinations and that's enough for 
the entire alphabet. From there, 
you can write anything. 

So now you've got all this material 
in the computer represented by ones 
and zeros. But you still have to do 
something with it. That's where George 
Boole comes in. 

Boole was an Englishman who lived 
during Babbage' 8 lifetime. Boole 
lived from 1815 to 1864, and Babbage 
did most of his workfrom 1812 to 1842, 
but I haven't come across any evidence 
that the two knew each other. 

Boole developed something called 
Boolean algebra, which is really more 
symbolic logic than algebra. He 
also was one of the first people who 
argued that logic should be. a branch 
of mathematics, not philosophy, and he 
certainly had some good reasons for 
it. Today youfind a fair number of 
philosophy majors working with com- 
puters, and it '8 not as odd as it first 
sounds. The computers work on the 
same principles of logic that philoso- 
phy departments teach. 

Boolean algebra is the type of logic 
where you have those little puzzles 
that look as if they came out of 
algebra books, such as "If A is true, 
B is not." 

Nobody found much practical use for 
this until this century when a man 
named Claude Shannon was working on 
his master's thesis, and discovered 
you can change these logical statements 
into sets of ones and zeros and let 
"he computer use the rules of Boolean 

Igebra to make its own decisions about 

oat to do next. This is the sort of 
logic that should tell the computer it 
doesn't have to send you a bill if 
you don't owe the store ay money. The 
computers that haven' t been programmed 
very well are the ones that send you 
a bill, anyway. 



BOX 1579, PALO ALTO CA 94302 

They say one way to tell If you'd make 
a good computer programmer Is to take a 
puzzle like this one from Litton Indus- 
tries. If you can figure this out and 
think it's fun, you'd probably make a £ 
good programmer. If you're ready to 
throw up your hands in despair, you'd 
better stay away from programming. 

"If Sara shouldn't, then Wanda 
would. It is impossible that the state- 
ments: 'Sara should* and 'Camille 
couldn't 1 can both be true at the same 
time. If Wanda could, then Sara should 
and Camille could. Therefore Camille 
could. Is this conclusion valid?" 

Now that you know whether you should 
be a programmer or not, we'll go on. 

Another name you hear a lot is" that 
of John Von Neumann. He's the one who 
figured out that you could put the 
entire program into binary form. Just 
why he decided to do this I'm not sure 
because if there was anybody who didn't 
need computers, it was Von Neumann. 
One story about him is that one of his 
fellow researchers had stayed up until 
4:30 a.m. doing five problems with a 
desk calculator. Then he decided to 
play a trick on Von Neumann. Von Neu- 
mann came in the next morning, and his 
friend asked for help in solving the 
problems. In five minutes, Von Neumann 
had worked out four of the problems in 
his head. The other person, who still 
didn't say he already knew the answers, 
then announced the fifth answer. Von 
Neumann apparently was quite perturbed 
that someone could figure out a better 
solution to the problem than he could 
until i-h»y told him what was going on. 

The early computers worked on 
vacuum tubes, and that soon got to be a 
problem. Vacuum tubes get very hot and 
they burn out. They're like a light 
bulb. It doe an' t matter how good they 
are; sooner or later they're going to 
burn out. The ENIAC, which was the 
first totally electronic computer, had 
17,000 vacuum tubes. And it wasn't long 
before computers were getting so big 
that if you made them any bigger, it 
would take 24 hours a day Just to re- 
place the vacuum tubes that had burned 

Fortunately, about this time— in 
1947 to be exact — the transistor was 
invented. Transistors do the same thing 
as vacuum tubes, but they're much tinier 
and they never burn out. It is possible 
to destroy a transistor by dropping it 
or by running too much current through 
It, but you really have to work at it. 

This solved a lot of problems, but 
it basically got computers down from 


the size of a small house to about 
the size of a amal living room. You 
oould,by the late 1950s, use tran- 
sistors to make radios small enough 
to pick them up and carry them around 
with you, but computers still needed 
too many transistors to be very 

Ejy this time, you had to do your 
work under a microscope, but scientists 
kept on working. In the lie 1960s, 
something called an Integrated cir- 
cuit was produced. Dr. Robert N. 
Noyce, the president of Intel Corp. in 
Santa Clara, is generally credited 
with being one of the co- discoverers 
of it. What the Integrated circuit 
means is that you can put the entire 
electronic circuit on a single piece 
of material, usually silicon. 

These are so minute that it's hard to 
believe. This is a silicon chip. 
What's even more amazing is the fact 
that it's Just the little gray spot 
in the middle that does all the compu- 
tations. The rest is here because you 
can't connect wires to something as 
small as the chip. But these gold 
lines eventually oonnect with 14 tie 
hairlike silver wires that lead into 
the silicon. I don't know how many 
transistors are on this particular 
chip, but some have 100,000* It may 
soon be possible to put a million 
transistors on something this size. 

To give you an example of the dif- 
ferences in size the integrated cir- 
cuit has meant, it's possible to put 
the entire UNIVAC computer, which was 
the first commercial computer, on one 
of these. 

If your family was one of the first 
in the neighborhood to have a tele- 
vision, you may remember the UNIVAC 
which was a guest of sorts of "People 
Are Funny." It used to spew all its 
cards out in front of the camera and 
Art Linkletter would pick them up and 
read off the names of two people who 
would get to go on a blind date 

The UNIVAC is now in the Smithsonian 
Institution, and it's the little chips 
like these that are taking over the 
world. Probably every piece of machlnerj 
at the Faire here depends on these 
silicon chips for its operations. 

The silicon chip starts wth a very 
unexotic raw material: sand. A 
shovelful of sand can supply the basic 
raw material for an entire computer. 

Silicon companies take sand and pure 
silicon "seeds," which are sold by 
only three companies in the world. They 

22 BOX 1 579, PALO ALTO CA 94302 

use these and "grow" cylinders of 
material which look like shiny, gray 

Once you get the silicon, the hard 
part comes. You have to put the trans- 
istors on it. The way you do it is 
sort of a cross between batik and 

In batik printing, you first draw 
the design on a piece of cloth. Then 
you decide which parts you want to turn 
out a particular color and cover every- 
thing else with wax. You dip it in 
a vat of dye, let it dry and scrape the 
wax off. The next time you cover 
everything but another color with wax, 
dye it again, scrape the wax off again, 
and keep on going until the picture is 

To make a silicon chip you do al-? 
most the same thing except that lay- 
ers of silicon oxide take the place 
of the wax and tiny lines of metal 
form the picture. The lines are so 
small and so thin that they are put 
on the chip through a photographic 
process in much the same way that 
shining a light through a negative 
produces a picture on a sheet of 
photographic paper. In this case, the 
negative is called a photomask. 

To make a photomask, you need a 
master diagram of the chip. These 
drawings start out several feet square 
and are reduced to the size of a 
chip, again through a photographic 
process. The masks, which are made 
out of glass, are made from these 
drawings • 

Each chip needs eight to ten dif- 
ferent photomasks, but there are 90 
to 100 steps involved by the time 
the chip is cleaned and new layers 
of oxide are formed on it and scraped 
off between photography sessions. 
And people are already working on ways 
to eliminate the photomasks and write 
the diagrams directly on the chip. 
So far, though, the machines that do 
this cost over a million dollars. 

All along the way, it's a very del- 
icate process. People who make chips 
don't even let you take pencils inside 
the laboratory because they produce 
dust when they write on paper. The 
water used to wash the chips between 
the different processes has to be so 
pure that companies sometimes have 
their own water purification plants. 

One firm — Monolithic Memories in 
Sunnyvale — says its water is 100 times 

I purer than distilled water. And when 
the plant finishes with it, it's still 
100 times purer than tho regular city 


When you start talking about chips 
this size, you find that computers 
are almost becoming self-perpetuating. 
It would be impossible to make them 
this small if you didn't already have 
computers to help do it. Computers 
are used to test the models of the 
circuits, they draw the layouts for 
the photomasks and thy control the 
manufacturing equipment. 

But theycan't do it all by themselves 
yet. When the whole thing*. is finished, 
that's when they call in the humans. 
The humans look through a microscope 
to check all of the circuitry on the 
tiny chips. The people at Monolithic 
Memories tell me that after a few 
weeks of training, people learn to 
check one in about a minute. The 
reason they need people is that there 
are so many different structures and 
so many differences in the size and 
the color of the lines that are 
still acceptable that there's no way 
to program a computer to remember 
them all. 

There's still not a computer that 
can remember as much as even the 
average human brain. 



BOX 1579, PALO ALTO CA 94302 


Robert Moody 

2233 El Camino Real, Palo Alto, Calif. 94306 

Phone: Home (408) 225-3341, Work (415) 327-8080 

I Introduction 

Computers are now within everyone's reach! 
Whether you are a housewife, small businessman, 
student, professional, musician, or in real es- 
tate — computers are being made and sold at 
prices you can afford, and they will set you 
free in ways you never dreamed of! A personal 
computer will open possibilities; it will al- 

use for communication between you and your com- 

Bit : The smallest unit of measure in a computer 
word; several bits make up a byte, or computer 

Bug : The cause of a malfunction, usually in a 
program. They're called "bugs" because they can 
be hard to find. 

most certainly change your lifestyle. 

Most people, when they visualize a computer, Byte : The space which a letter or digit (one 
they think of monstrous machines that tower over character) takes up in a computer. Space in a 
us, seeing all and doing all. It's really not computer is measured in bytes. A megabyte is a 
that way. Everybody has been communicating with million bytes, 
a computer in one way or another most of their 
adult lives and not really known it. For ex- 
ample, all your tax returns are processed by a 

computer, most amjor department stores handle 
their buying, stocking, and billing by computer. 
All your credit card buying is handled by com- 
puter, every check you write is processed by 
computer. Dentist, doctor, hospital, gas, elec- 
tric, phone bills are handled by computer. Why!!! 

Take a city the size of San Jose for inst- 
ance. 3ust think of the tremendous amount of 
manpower it would "take" to try" and post all the 
checks that were written in one day, or the 
amount of phone calls to be logged in one day. 
Handling that amount of information — having 

Computer : A machine for handling repetitive in- 
formation. Basically it can calculate, compare, 
alter, send and receive information very rapidly. 

Core : See memory. 

CRT: Your computer's "TV screen", showing you 
what's IN there. The CRT is your computer's way 
of talking to you. It is also something refer- 
red to as a display unit, or terminal. 

Data : The information that gets WORKED OVER, 
when your program runs. Data is all the in- 
formation you have your computer use, everything 
that is sent into your computer to store and 

that kind of computing power - is at your finger Disk . ft ^^ storage dBKlicBf eithsr fi oppy or 
tips today. 

The most important think that I want to 
convey to you is: computers are not just for 

geniuses!!! You don't have to be a special 
gifted person to own or operate one. 

This presentation, I hope, will enlighten 
you a little as to what are these things, 
personal computers, how you can talk and ask 
questions about them, what makes them up, and 
what you can do with one. 

II Buzz Words 

As you know, all types of industries have 
their own language that they use. We as a group 
and a new up-coming industry, are no exception. 
As you scan down the second page of the handout 
I have given you, there is a short list of these 
buzz words, and a simple explanation of each. 

BASIC : Beginners All-purpose S_ymbol instruc- 
tional Code: a mode of language that you will 

hard disk. 

Display : Same as CRT. 

Floppy Disk : A mass storage device, which uses 
a flexible platter to store a large amount of 

Fortran : Another type of computer language. 

Hard Copy : Computer output ON PAPER, for per- 
manent storage. 

Hard Disk : Much like a floppy, a hard disk 
stores a tremendous amount of information, but 
its platter is much larger and not so portable. 

Input : The information that goes IN to your 
computer system. The computer's "food". 

Interface : A connector that "translates" be- 
tween two parts of a system. You generally need 
one interface for each peripheral , to hco'k it to 
your computer. 



BOX 1579, PALO ALTO CA 94302 

Mass Storage ; Any way of keeping a lot of in- 
formation OUTSIDE your computer, but available 
to it. This is your computer's "memroy". Most 
common kinds of mass storage are tape and disk . 

Micro, Microcomputer, or Microprocessor ; Same 
as computer. The "micro" came in when we 
learned to make them physically tiny — they are 
about the size of pencil erasers. 

Output ; What your computer system produces. 

Peripherals ; The devices attached to your com- 
puter, such as the display , keyboard , printer , 

RAM ; Random Access Memory. This storage device 
is used by your computer to change the data you 
have put into it, then it is transferred to a 
mass storage device. 

Storage ; The part of your system that remembers 
information, as opposed to the parts that 

Software ; A list of instructions to the com- 
puter, telling it what to do and when to do it. 

Terminal ; A unit for conversing (input or out- 
put) with your computer. It has a keyboard, 
plus Display or Print-out. 

TUT; "Television typewriter". A keyboard and 
electronics specially designed to turn your TV 
into a TERMINAL. 

Ill what Makes Up a Computer? 

In your handout you will find a block dia- 
gram of a computing system. As you notice, 
there are not many modules or components that 
are needed to do the job. The technology today 
has made this possible with microelectronics. 
With this tremendous reduction in physical size 
of transistors and diodes it enables us to com- 
pact a large amount in a vary small -area. Also, 
the power requirement is small as well. 

The first and most important module is the 
CPU, or Central Processing Unit. This is the 
brain of the personal computer. The CPU does 
all the actual work of calculation, comparison, 
alteration, receiving and sending data. . Every- 
thing else that we attach is a function from the 

There is a wide variety of CPU's on the 
marketplace today. Depending on who you talk 
to, one is better than the other. Don't let 
this discourage you now; all that you need to 
know at this point is that they work. You can 
get into particulars later. 

The CPU cannot operate by itself; along 
with it you need some memory. RAM or Random 
Access Memory does the job. This portion of the 
computer allows the CPU to activate a section of 
the data at one time. For instance, the RAM 
could be compared to an active filing system. 
The data is stored in some kind of order and the 
CPU only pulls the data it needs and keeps the 

rest for later. 

Now that the CPU has changed, calculated, 
or compared this data, it needs to store it 
someplace. This is where mass storage comes in. 
Here again there is a wide variety and different 
types to choose from, but basically it is com- 
pared to an inactive file. This file holds much 
more data than the RAM in your computer does. 
You as a computer operator put in and take out 
these files at will, making one active and an- 
other inactive. 

To be able to accomplish the task of moving 
all this data around, you have to be able to 
talk to your computer and have it talk back to 
you. A peripheral device called a Terminal is 
necessary. This is usually made up of a key- 
board, much like a typewriter keyboard, and a 
display unit. The Terminal is attached to your 
computer through an interface device. There 
are two basic types; serial and parallel. 

The serial I/O, or input-output, interface 
takes the data you are typing in and sends it 
one bit at a time to the RAM. Parallel, on the 
other hand, sends all the bits that make up the 
computer word or byte, and puts them all in at 
one time. This might seem very confusing to 
you, but all that is necessary to know is what 
kind of interface is incorporated on the peri- 
pheral you are attaching to your computer, and 
match it up with the proper I/O device. 

Now that I have thoroughly confused you, 
we will push onward. 

IV Programming 

Now that all this hardware has been 
assembled, we have all 'the physical things that 
are needed, but there is something that is 
lacking; software. What this thing does is 
allow the computer to try and make some sense 
out of what you are trying to tell it, and vice 
versa, have you understand what it is trying to 
tell you. Fundamentally there are two types: 
systems and applications. 

Systems software is what the computer uses 
as its language. It takes the information in, 
in this special language, and acts on it. What 
you do as an operator is use this language to 
develope an application. 

There is a lot more to programming than 
just a couple of sentences I have used to de- 
scribe it. I could ramble on about all the 
different types of systems software, but I have 
to spend more time on the biggest question that 
I'm asked, and that is "I would like to have 
one of these things called personal computers, 
but what the hell do I do with it?" 

V What Can I Do With It? 

The most asked question that I receive is: 
"what can I do with this thing, now that I have 
it?" The uses for personal computers are end- 
less; the list below only shows a few uses — 
let your imagination go and you will see! 



BOX 1579, PALO ALTO CA 94302 


Accounting and billing 

Reducing the physical 
SIZE of FILES (custo- 
mer files, corres- 
pondence, product 
files, etc.) - No 
more file cabinets ! ! ! 
Playing games 
Engineering design-aid 

Maintaining "tickle- 
files" (calendar- 
reminder systems) 
Filling out checks 
Playing the stock market 

Remembering all trans- 
Making and keeping card 

"Simulating" results of 
one decision versus 
another, so you can 
see their effects 

Inventory management 
Routine correspondence 

and form letters 
Filling out forms 
Calculation of all 

Receiving and placing 

phone calls 
MATCHING any informa- 
tion with any other 
Polls and surveys 
Solving problems 
Printing out results 
Maintaining LISTS, 
Sales analysis 
Travel and route 


Lists — Computers are lovely at working with 
lists. The nice thing is that you can change 
anything in the list — even one letter "B" for 
example — without affecting the rest. It's 
like a stored-away blackboard. As items become 
obsolete — the way they're always doing in a 
shopping list, say — you just tell your com- 
puter to "drop" them from the list. Your com- 
puter does the rest: YOU DON'T HAVE TO KEEP 


This makes your computer a powerful tool 
for such unexpected things as TEXT EDITING. You 
can work a whole manuscript over without having 
to erase or "fix" anything, physically. The 
time savings alone are immense — in writing, ed- 
iting, and composition! But imagine USING NO 
PAPER until you've got the final version ! 

With all this, your Personal computer adds 
Improvement: Your computer will do FOR you what 
you're already doing, in a fraction of the time. 
Expansion: You do things with it you never 
could have before. 

Improvement — If you; re in mail-order, small 
business, or an office profession, a "micro" 
computer is for YOU! It can keep your mailing, 
li 8 t8 and print address labels at a rate of four 

per second, all accurate and up-to-date. It can 
store all your records on products, sales, cust- 
omers and ad results. It will compare or change 
any part of these you'd like, without touching 
the rest. In other words, it gives you SALES 
ANALYSIS and MARKET PROFILES which only "multi- 
millioners" have had until now! 

Your computer keeps inventories . 

It does accounting and billing of all typea 

It does order entry (live). 

It can hold and index your " own private 
library" of whatever information interests you. 

For example, it can help in health care by 
keeping track of each patient's medical history 
and alerting you to patterns you might have mis- 
sed. Or give you a quick way to look up the 
newest therapies by what they remedy — you type 
in the condition and your computer gives you 
back a list of indicated drugs or treatments. 

If you are a pharmacist: you may presently 
be keeping "card files" of each customer's 
contra-indications and other medication history. 
With a computer, you'd no longer have to look 
each of these up for each prescription! You'd 
just type in the customer's name and Rx, and UP 
would come the pertinent information on him! It 
would be shown on a little TV screen — just 
like at the airlines. You could also tell your 
computer to remember a list of "potentiating" 
drug combinations and warn you if it detects one 
in a customer's combination. YOU DON'T HAVE TO 

Your computer can run FLEXIBLE FORM LETTERS 
for you! You tell it to change this word or 
that, add or delete a paragraph, date a letter 
next Monday, and address it to, say, all the 
people in your "Best Prospects" file. OUT will 
come a stack of letters, individualized and 
ready to mail ! Because of "the '""add-on"" equip- 
ment which adapts computers for particular jobs, 
we arrive at your computer's greatest promise: 

Expansion — You're about to see what new types 
of business are possible with a computer running 
in your own home! These are all high-profit, 
high-service businesses — yet they take little 
SPACE or TIME. Best of all, they're VERY low in 
STRESS! Because, let's face it, one of the 
chief causes of STRESS seems to be EMPLOYMENT. 
It's hard on both worker and boss . So why be 
either? GET A COMPUTER. 

Here are some new businesses which one per- 
son alone can set up with a computer. The first 
are all based on one thing computers do very 
well: INFORMATION-MATCHING! Basically this 
gives you a CLASSIFIED AD SERVICE - either gen- 
eral or very special (loke clothing only). And 
it can be "live", available right over the phone 
OPPORTUNITY! . It works like this: people call 
to say they want this or have that, and you tell 
'em who has or wants it. You do this by signal- 
ing your computer what's wanted. In a few 



BOX 1579, PALO ALTO CA 94302 

seconds it gives you a LIST of people who hav/e 
or want that thing! You read that to your 

Your computer will display any desired de- 
tails you've given it in the past, as part of 
the list. The point is, YOU don't have to 
REMEMBER it all anymore. So your "data bank" 
can be much larger than you could ever handle 
mentally, and you're able to give far broader 
service. At the same time you're freed up for 
creative work! 

Moreover, some phone companies are learning 
to cooperate by "On-Call Billing" of calls to 
your Service number. The caller is simply 
charges a higher rate for calling that number. 
These charges are then forwarded or credited by 
the Phone Company directly to you. In other 
words, as a business you get paid when you're 
called . 

Where On-Call Billing isn't available yet, 
you can tie-in your fee with transactions, like 
the New York "BUYLINES" do. This is more work, 
but that's what your computer does anyhow. 

Some new satellites are making phone lines 
unnecessary, and getting clearer transmission to 
boot. Your "phone system" can become very pow- 
erful, with help like this. By "being in" with 
a system ahead of time, you can make the most of 

New Business — 

Real Estate - with up to the minute list- 
ings over a wide area. 

Dating Service - getting people together. 
With speedy, detailed descriptions that YOU 
don't have to look up! (Your computer will spot 
the most "ideal matches".) 

Apartment and Building Rental Service - 
where subscribers can find places for rent , or 
list places they have. There's a need for this 
kind of service, since real estate agencies are 
more interested in SALES. If you have an agency 
however, a computer lets you expand painlessly 
into rentals. With a computer you can make your 
service "long distance" too, using classifieds 
from other cities. 

Swappinq Service - You put people in touch 
with each other, who need and have things they 
want to trade. Instead of money, they give you 
"due bills" which you charge 10$6 each on. These 
act as "certificates" for exchanging the goods 
or service. For example, you know a dentist who 
wants a used car; and a car dealer who wants his 
house painted. Your COMPUTER completes the cir- 
cle: it displays the name of a painter who 
needs dental work! You have due-bills from each 
already, and just put the people in touch. And 
you get paid with each due-bill — so it doesn't 
matter to you whether they go ahead with the 
swap or not. 

Equipment Rental Service - here your com- 
puter merely keeps tabs on who has equipment 
they are willing to rent, and of who wants to 

rent some. Again, YOU DON'T HAVE TO STOCK ANY- 
THING. Furniture, housegoods, and even cars can 
be rented. 

Shopper's "Where To Find It" Service - this 
is like a detailed "Department Store Directory" 
available by phone. Your computer keeps track 
of what stores carry which items, and prices too 
if you wish! You can also notify of specials 
and sales. 

Instant Babysitting - you run a "register" 
of babysitters who are available on short notice 
This can be used for many other types of jobs 
and services too, of course. 

"Phone Rummage" Sales - Like a rural ver- 
sion of the What's For Sale service. The big 
items can even be cataloged into your system, so 
callers can ask, say, for a piano and you'll 
know where one's for sale. This can be applied 
to flea markets, too. 

Clothing Clearinghouse - this answers an 
acute need. Everybody — especially females 
and children — had a predicament until now: 
they need different nice things to wear, and 
they have good clothes they don't want. Yet 
they're unable to sell or find a used thing ! 
(if you'd seen the pennies a store offers you 
for a perfectly good $300 gown you've worn once, 
you'd see why!) This is a natural for your com- 
puter. You just keep pumping it full of the 
clothing descriptions and clients' names and 
phone numbers, and your computer will find every 
match there is! - even down to desired color and 
size! After you've brought two parties together 
they can work out any terms they wish about the 
clothing. You get paid by the phone company, or 
by clients for the numbers you give them. You 
need NO SPACE and NO MERCHANDISE to conduct this 

Now for some businesses that are NOT of the 
"INFORMATION MATCHING" type. Notice that these 
still take almost NO SPACE to operate. 

Service Bureau - doing small business work 
for other people or firms. Accounting, billing 
payroll, mailings, correspondence, taxes and in- 
ventory are top candidates. You can do for 
others , anything -we've named so far. 

Answering Service - fully automated. We 
think that with a little ingenuity you could run 
a whole answering service using nothing but a 
phone and your computer. I mean, go off to the 
Health Club while your computer routes messages 
all day! 

Calling Service - Your computer, properly 
rigged, can do much of your routine phone cal- 
ling for you. You can even tell it to call 
someone from across the room! (And IT will 
remember the number - all you have to know is 
the name). Some people will use a computer to 
place automatic "buy-sell" orders on the stock 
market or commodities. You could be earning in 
two places at once! You'd be at your job or 
business, while your computer is playing the 



BOX 1579, PALO ALTO CA 94302 

stocks for you! 

Horoscopes - your computer can cast horo- 
scopes, and then print them out for you to sell. 
It's able to mix "individual" information such 
as birthdate and place, with the "mass" informa- 
tion about astrological types. (You choose and 
give it the "mass" information ahead of time. 
It can then print out horoscopes in as many cop- 
ies as you wish, with even the peoples* names 
and addresses on them for mailing. You see such 
horoscopes advertised nationally. Those ads net 
hundreds of thousands of dollars ! 

Tarot Readings - this is the same idea as 
horoscopes, only the "general" information you 
give your computer beforehand is from books on 
the Tarot instead of astrology. Your computer 
would "draw" the cards for you! Tarot readings 
are just getting popular, and more novel to the 
public than horoscopes. You could see them 


Writing Service - your computer is the best 
Assistant Editor anyone ever had. How it saves 
TIME! And PAPER ! More on this in a moment. 

Classified Bulletin - publish it monthly or 
bi-weekly. It has nothing but classified ads of 
all kinds. The ads are run FREE and people pay 
you when there's a sale. This sounds risky, but 
for many reasons, IT WORKS. The Classified Bul- 
letin is a much better deal for people than the 
newspaper — all you need is adequate coverage. 
People's ads run longer, run free , and cover 
better geographical area. You could be running 
a "live" classified service and a printed ver- 
sion, at the same time. Each serves as public- 
ity for the other. You could also support your 
classified bulletin with paid ads from local 

-Other new businesses possible, with your 


"Local Events Calendar 

Reservations Service 

Opinion Polls and 

Lottiers and other 
Fortune Games 

Betting Systems - win- 
ning at horses, black- 
jack, etc. It can be 
done, in fact it IS. 
And there are some 
books on it. 

Gold Handicapping 

News Service 


Problem Solving for 

Special Interest Clubs 

Selling Time on your 
computer. You charge 
other people for 
using it. 

Your Own Publication - 
magazine, classified, 
or newsletter, edit- 
ing with the computer. 

You're getting an idea of your computer's 
business uses. We bet you'd also like to know 

How a Computer Can Help You in Your Home 

Now with all of these uses, you must be 
starting to guess what one of these "mighty 
midgets" can do in your home! First: you can 
have everything controlled automatically, whe- 
ther you're there or not. 

Your home computer can literally become 
your "private secretary". Make you feel like 
the President, as it keeps track of notes, dates, 

reminders, appointments... It will alert you at 
the right time — give you a message that 
"school's letting out early today," or that So- 
and-so's coming to visit, or that you have an 
appointment downtown at 2:00. Keep lists, bal- 
ance your checkbook, be a library, manage your 
finances and answer the door. YOUR MIND GETS 
UNBURDENED — a computer's really useful for thati 
Maybe even prevents wrinkles , if those come from 
feeling harassed!!! 

Right now, you'll "tell" your computer what 
you want by typing messages on its keyboard. 
Don't worry how it "understands" them. It does. 
You just type your instructions, and it'll do 
the rest. 

Now you'll even be able to SPEAK your com- 
mands to it, and it will carry them out! For 
example, there's an alarm clock on the market 
right now that stops ringing if you yell at it. 
(If you just groan or talk, it rings less.) And 
it recogni es YOUR voice — no one else's! 

Any electrical device can be set up to do 
this. We've always liked the idea of LIGHTS 
that keep themselves on only when someone's 
around . (You wouldn't have to talk to them, 
they'd know you were there by your temperature.) 

As each of these technical devices comes 
onto the sales market, you'll be ready. You can 
just ADD them to your system. 

Your computer can also help with home pro- 
tection . It's hard to beat a trained attack dog 
of course. But he can't call the Fire Depart- 
ment if he smells smoke, or signal your "beeper" 
when the baby wakes up. And you don't have to 
ualk your computer. For burglary, as a matter 
of fact, a little machine is planned that you 
attach to your door. It's activated while you- 
.. re. .awayj. to SOUND like _a_ fero ious dog roaring 
and hitting the door, if anyone starts tampering 
there. Your computer can do even better, espec- 
ially if you have a Hi-Fi set. 

It could even be wired to water plants or 
feed fish. 

AT HOME AS IN BUSINESS, your computer's 
ALL KINDS OF PLANNING. You can do budgets, 
schedules, and calendars, party management, in- 
ventory, and keep your own "mailing lists" and 
phone directory. You can do super MENU-PLANNING 
with a computer, too. It tells you what's in 
the refrigerator, what you're out of, what needs 
using up, and what you'll need to buy for a cer- 
tain recipe. Or it tells you what recipes you 
can do with what's on hand. Frankly, all this 
would be a big order for your little computer at 
present. But it's coming (Also, the longer 
you spend with your system, the more versatile 
it gets.) 

Remember how easily your computer handles 
LISTS. Well, you can add or delete items all 
week, and review it anytime you want. Only 
those items you changed will be different. And 
the rest of the time your lists are tucked away- 
safe and CLUTTER FREE! 



BOX 1579, PALO ALTO CA 94302 

■ Now more on TEXT EDITING, because it's so 
I ue.ll suited to WORK-AT-HOME! (The sources of 
I editing work to do at home are so varied that we 
I won't go into them here. The commonest is TYP- 

■ ing, of course.) 

I You type your material into the computer, 
| which "stores" it in its memory. Then you call 

up the material section by section, and work it 
over. YOU USE NO PAPER. You don't have to 
erase, or "fix", or buy correction fluid. You 
don't have to make carbons. Any change you call 
for is made instantaneously. After the work is 
"PRINT" button. THEN your system spits out fin- 
ished pages. The newest systems do this a whole 
page at a time , and silently like a Xerox. 

Something else your computer can do for you 


We don't know if you're into indexing or 
not. Some are, some aren't. If you have a lot 
of ideas or keep a diary or want to sav/e clip- 
pings and the like, thouch, INDEXING is for you. 

flair — just like programming. So, you mightn't 
have guessed what fun for YOU lies hidden in a 

NOW. . . your home computer can act as Door- 
man, Guard, Secretary, Librarian, Mother's 
Helper, Consultant, New Business Partner, Com- 
panion, and just general Unpaid Labor Force. 

But bet you never thought of it as an 

You're hearing a lot about "TV Games" for 
this Christmas. They're just "little black 
boxes" you hook up to your TV. You play games 
with them through a little keyboard, using the 
TV screen as a "display". 

Know what those "little black boxes" are? 
They're microcomputers, that's what! Someone 
else has programmed them, is all. 


One of personal computing's greatest pro- 
mises is in MUSIC. Imagine yourself going to 
your piano or organ, and playing as slowly as 
you wish ! ...Maybe one finger at a time, or 

It makes a compact filing system, ready to go — making it up as you go. All the while, your 

and your computer's the one to do it! It will 
keep track of all the headings, key words, page 
or issue numbers, and the alphabetizing of 
everything. Here's how it works, basically: 
Say you have a clipping or an idea you want to 
"index". Say it's about a car that runs on sun- 
power. You just tell your computer the key 
words it's about - like FUELLESS CAR - and 
then tell your computer where you've put the 
writeup on it. Like, "file of October '76, Item 
Number 4." 

Then, say a few months later you want to 
look up the article again. You just tell your 
computer the KEY WORDS - type in FUELLESS, CAR - 
and your computer tells you exactly where to 

computer is "listening". It's remembering 
every note. (And no-one else has to hear you., 
you can wear headphones!) 

When you're finished with your "composit- 
ion", you tell your computer "play it back - to 
tempo". Your computer has the piece played 
back, as fast as you want it to go . 

There you are — you've made your own player 
piano or "player organ"! And what's it playing? 
YOUR piece! Speak of ENTERTAINMENT—!! This is 
an indescribable experience, if you've always 
longed to make music, but never got the chance. 

Present-day organs are moving in this di- 

look for the writeup! (Yo< might have collected rection, but they still require a lot of skill . 

several items on this subject. Your computer 
would then tell you about all of them.) It's 
like your own private library. 

If you'd collected a lot of other material 
on cars, then you could see it all by just typ- 
ing in "CAR". Or maybe you're interested in 
anythinq that runs without fuel. Then you'd 
just type in FUELLESS, and get back more than 
cars. See? 

INDEXING is like a great scrabble game. 
It also tends to make you smarter, for some 
reason. (Certain studies prove this.) Maybe 
that's because indexing develops both your 
logic and fine judgment at the same time. 

There are several ways in which indexing 
can be built into a business . Cutting and sel- 
ling newspaper clippings is who's likely to pay 
most for what subjects. Likewise, which publi- 
cation. Your computer does a superb job of 
this. Another business possibility is abstract- 
ing the literature in a given field. The ab- 
stract can be sold, and so can the indexes! 

Some people just "take" to indexing - it's 
another one of those areas for unpredicted 

You have to be able to "keep up with the 
rhythm", for example. With your own computer, 
you won't have to: the rhythm keeps up with 

How Your Organization Can Use a Computer 

A computer can be a real asset for a SMALL 
ORGANIZATION. It's indispensible in event 
planning, maintaining calendars and schedules, 
bookkeeping, assigning and tracking committee 
work, reporting, calculations, issuing ads and 
announcements, doing memos, and keeping the 

These can all be INDEXED to let you find 
the exact material you were thinking about, 
stored in the computer four years ago! 

Your computer is also the center of atten- 
tion at social events, as it handles ticketing, 
door prizes, drawings and countless crowd-parti- 
cipation games. Imagine the attraction THAT is 
with you as the sponsor! 


Speaking of community, what if you took 



BOX 1579, PALO ALTO CA 94302 

computer like this to SCHOOL? It could make you 
a hero. And since you programmed it, of course, 
it's ready to do whatever you planned. 

As a teacher you could use it as a TUTORIAL 
AID and to do INSTANT GRADING, for example. 
Perhaps above all, you could use it to TEACH 
PROGRAMMING. And the students would have to 
write programs to DO something, so they start 
coming up with interesting applications. (Again 
programming boosts mathematical ability — it's 
not the other way around, as often supposed.) 
The kids find themselves exposed to possibili- 
ties like all these we're talking about. They 
catch on quick what to do with them! 

Perhaps one of the most important things 
kids can discover is how to earn a living in a 
way that feels creative to them. And again we 
see the LITTLE computers playing a BIG part in 
exactly this discovery! 

Meanwhile, the little computer can help 
with HOMEWORK. And it can feature in all sorts 
of projects and reports, ones that get writeups 
in local papers. This often happens with stud- 
ent computer projects, because they tend to be 
so original. Also, one student can do the work 
of many, this way. And likewise, many scien- 
tific projects become possible with a computer, 
which were impossible before. 

VI The Next Step 

Where do you go from here? Well, my sug- 
gestion is more reading. Take what I have given 
you today and expand on it. There are a number 
of books out in publication now that will enable 
you to understand better all this that is going 
on and in much more expanded form than I have 

Make use of your local computer store! The 
owners and operators are more than willing to 
give you more of an understanding and help you 
over the hard bumps of decision-making process 
of which one to buy and what to hook up to it. 
Also look into classes at your local high 
school and college. There are any number of 
classes to take to further your knowledge in the 


As a final note, remember: Computers are 
not just for geniuses, you don't have to be a 
special gifted person to own or operate one. It 
is a tool that you can use, that will help in 
every day life. 



BOX 1579, PALO ALTO CA 94302 


Dr. Robert Sliding 

Research Director for the Digital Group 

Box 6528 

Denver, Colorado 80206 

Abstract: "Handicapped" can refer to either physically impaired 
or mentally impaired. There is a need for individual hobbyists to be- 
come involved in aiding these handicapped people. Present technology 
supports a large number of exciting applications. Hobbyist magazines 
are begging for articles on how to implement systems for the handicap- 
ped. An international organization of hobbyists interested in handi- 
capped applications was formed at PCC 77(Atlantic City Convention). 
For more information or to join contact: 

Computers for the Handicapped 

c/o Warren Dunning 

5939 Woodbine Ave. 

Philadelphia, PA 19131 


Reprinted from People's Computers 
1263 El Camino RmI 
Manlo Park, CA 94025 


g^ .i _<■»«.. *? 

for the 


By Tim Scully 

This article is more technical than many 
published in People's Computers, but we 
believe that the general discussion will 
be interesting, informative, and thought 
provoking to all, even those who choose 
to skip the program listings and discus- 

Tim Scully has been designing biofeed- 
back equipment and doing biofeedback 
research for many years. Tim is a Re- 
search Fellow of the Humanistic 
Psychology Institute; he is now working 
towards his doctorate in psychology. His 
dissertation project involves researching 
and developing biofeedback systems and 
techniques for use in drug rehabilitation. 

Tim is also teaching a computer class to 
fellow inmates at a Federal penitentiary. 
Although prison resources are scarce and 
he is not allowed to solicit donations, he 
is hopeful of somehow eventually acquir- 
ing a computer system for the prison. 

The potential of microcomputers as tools 
for the handicapped is enormous and 
exciting: we encourage dissemination of 
such information. For this reason we are 
making copies of this article available. 
To receive a reprint, send a stamped, 
self-addressed envelope (24t for business 
size, 35^ for 8% by 11 inch) to People's 

How would you communicate if you 
couldn't talk, didn't have the use of your 
hands, and could only somewhat control 
the movements of one knee? This is the 
problem which Robin, a young lady in 
her 20's has lived with all her life. She has 
cerebral palsy. 

I met Robin in 1976, and this is the story 
of how a microcomputer communication 

system came to be built for Robin. The 
general concepts applied in the develop- 
ment of Robin's communication system 
may prove helpful in the development of 
microcomputer systems for other handi- 
capped people. 

When I first met Robin, her communica- 
tion was accomplished by use of a word 
wheel. She could understand speech and 
she could read, but she needed help in 
'talking'. Her word wheel was made from 
an electric clock motor and a bicycle 
spoke, with the bicycle spoke attached 
where the second hand of a clock would 
normally be mounted. A sheet of card- 
board was mounted behind the spoke, 
with the letters of the alphabet on it, 
arranged in a circular pattern. The spoke 
pointed to the letters, one at a time, as 
tt r otate d. Robin could move her knee 
to one side and hit a kneeswitch mounted 
on her wheelchair, thus stopping the 
motor so that the spoke would freeze, 
pointing at the letter she had chosen. 

The spoke rotated at one revolution per 
minute, so spelling proceeded at about 
one letter per minute! The person Robin 
was conversing with often had to write 
the letters down, to keep from forgetting 
them, as a message slowly built up. 
To speed up the communication process, 
a few words were written next to each 
letter of the alphabet, so that when the 
spoke stopped it would point at a group 
of words as well as a letter. The person 
with whom she was conversing would 
have to guess which of these Robin 
intended. It took considerable patience to 
hold a conversation with Robin, and not 
very many people took the time:- 

When I first saw Robin's communication 
system, I thought of replacing her word 
wheel with a microcomputer and video 



BOX 1579, PALO ALTO CA 94302 

display, using a vocabulary of words 
stored in the computer's memory in place 
of the sheet of cardboard. A little over a 
year later, that system now exists and is 
being installed on Robin's wheelchair. 


The present system is an expansion of the 
word wheel concept which uses a TV 
display with 1 6 lines of text. The top line 
is reserved for the display of a 'menu' 
of items (words, letters of the alphabet, 
punctuation symbols or control codes) 
from which Robin can choose. The 
second line is kept blank and the bottom 
14 lines provide space for the display of a 
message of about 200 words. 

As items are displayed on the menu, 
Robin can choose one by hitting the 
kneeswitch mounted on her wheelchair. 
In some modes of operation several items 
will appear on the menu at once, in which 
case the item at the left is the current 
item, the one which can be selected by 
hitting the kneeswitch. 

On start-up, the system blanks the TV 
screen and then offers the SPELLING? 
mode by putting that word on the menu. 
This item remains on the menu for a 
time 'Tl' (an adjustable time delay). If 
the kneeswitch is hit during that time, 
the SPELLING? mode is entered, other- 
wise the next menu item is displayed: 
PUNCTUATION?. If that item isn't 
chosen either, after another delay equal 
to Tl, then the system will begin 
displaying the names of groups of words: 
TWO-YOURSELF, one group at a time. 
Each group of words contains about 
120 words in alphabetical order. The 
name of each group is made up from the 
first and last words in the group. 

If Robin doesn't pick any group of 
words, the computer then offers an 
ESCAPE? from the groups of words. If 
this isn't chosen, the names of the groups 
are offered again. If the ESCAPE? is 
chosen, the system returns to near the 
beginning of the program and offers 
SPELLING? again. This ESCAPE? to the 
beginning is offered from every mode of 
system operation. 

If Robin does pick a group of words, 
HIGH -LOT for example, then the names 
of subgroups in that group begin being 
displayed, one at a time: HIGH -HONOR, 
ESCAPE?. If Robin picks a subgroup, 
such as LEAVE- LIE, then the words in 
that subgroup are displayed across the 
top line of the TV, with two spaces 
between each word: 

If Robin hits the switch at this moment, 
LEAVE will be transferred down to the 
first available space in the message 
area of the TV screen and the menu 
will begin all over again by offering 
SPELLING?. If the first word, LEAVE, 
isn't chosen, then after the usual time 
delay Tl, the list of words on the menu 
will shift one to the left, so that LED is 
on the extreme left and it becomes the 
current item. This process continues 
until a word is chosen or until the end 
of the subgroup, LIE. If LIE isn't chosen, 
ESCAPE? is offered, and if it isn't 
chosen, the complete list of 1 1 words in 
the subgroup is displayed across the menu 
and the cycle begins again. 

By this system of groups of words, 
subgroups, and finally words, it is 
possible for Robin to look through a list 
of 1200 words in a short time, find the 
one she wants and add it to a message she 
is assembling on the TV screen. The 
computer automatically adds a space 
after each word chosen, so it isn't 
necessary for Robin to worry about 
spacing between words— she can just 
choose one word after another. All letters 
and words are upper case, so she doesn't 
have to shift. 

When a sentence is complete, and when 
she wants punctuation symbols, Robin 
can select the PUNCTUATION? mode. 
The first item offered on entering this 
mode is CONTROL? and if that isn't 
chosen, then after the usual time delay, 
the punctuation symbols will be spread 
across the menu in much the same way 
that the words in a subgroup were 
.'?;:!012...9#$ %&()* + - 

These items leave the screen at the left, 
one at a time, if they are not chosen. If 
one is chosen, the computer backspaces 
once (to undo the automatic spacing) and 
adds the chosen symbol to the message 
on the screen. Then the system starts over 
by offering SPELLING? again. 

The CONTROL? mode offers Robin a 
few useful commands, one at a time, if 
it is chosen: BACKSPACE?, ERASE 
SCREEN?, and NEXT LINE?. These 
control codes operate immediately if 
selected. Then the system starts over by 
offering SPELLING? again. 

The SPELLING? mode exists to allow 
Robin to spell words not found in the 
1200 word vocabulary stored in the com- 
puter's memory. To speed up the process 
of spelling, letters of the alphabefcare not 
offered in alphabetical order. Instead 
they are offered in the order of their 
probability of use in English. Except 
at the beginning of a word, the likelihood 
of a letter appearing in a word depends 
on the last letter chosen .t If we are in 
the middle of a word, and the last letter 
chosen was 'A', then the most likely 
next letter is 'E', the second most likely 
is 'B', etc. 

Robin's system has 27 different alphabets 
stored in it. The first alphabet has the 
letters organized so that those most likely 
to appear at the beginning of a word will 
be displayed first. This is the alphabet 
which appears when the SPELLING? 
mode is first entered. The letters are 
spread out along the menu line as usual, 
with the first offering on the left. If no 
letter has been chosen by the time all of 
them have moved off the screen to the 
left, the usual ESCAPE? offering is made 
and the alphabet redisplays. 

If a letter is chosen, it is added to the 
message area of the screen, and ESCAPE? 
is offered on the menu. If Robin decides 
to stay in the spelling mode, the 
computer then displays one of the 26 
remaining alphabets— which one is deter- 
mined by the letter she just chose. 
When she picks a letter from this new 

t Mr A Ross Eckler suggested the bigram spel- 
ling scheme used in Robin's system. He supplied 
me with letter use frequency tables which he 
credited to F Pratt, Secret and Urgent: The 
Story of Codes and Ciphers, Blue Ribbon 
Books, 1942 pp 258-259. 



BOX 1579, PALO ALTO CA 94302 

alphabet, it is added to the message, 
immediately after the first letter (the 
system automatically backspaces to undo 
its automatic spacing). This process 
continues until she has completed 
spelling a word. Then she picks ESCAPE?, 
which returns her to the beginning of the 
program, which offers the SPELLING? 
mode, and a space is left after the word 
she has just completed. 

This spelling scheme allows comparatively 
rapid spelling of words because Robin 
only has to wait for a few letters to 
display before the one she wants is likely 
to become the current item. The 
automatic spacing also speeds up 

Now that we've looked at what Robin's 
system does, let's examine the hardware 
and software which do the work. 


Robin's system was designed around the 
special limitations of her situation and 
my own situation. I met Robin through 
a United States Probation Officer, who 
was supervising me while I was tempo- 
rarily free on appeal bond. I was waiting 
for the Court of Appeals to decide if it 
would uphold my conviction for conspir- 
acy to manufacture LSD (back in 1968 
and 1969). As it turned out, the Court 
did uphold my conviction, and I'm now 
serving a 10 year Federal prison term 
at McNeil Island Penitentiary in 

My personal problems limited the system 
design to the use of a commercially 
available computer kit because of the 
difficulty of sending materials into prison. 
Robin's family had only a limited budget, 
and Robin's capabilities formed the 
remaining design limits. 

In 1976, the budget we had (about 
$1,300) was just about enough to buy a 
computer kit with keyboard, cassette 
tape system, video monitor and 8K of 
memory, so this is the size system we 
planned on. The average word in English 
is about 5.5 characters long and we 
initially planned on a vocabulary of about 
1,000 words, which uses up 5,500 bytes 
of memory. This left about 2,500 bytes 
for the program to control the system 
together with storage for spelling and 
punctuation symbols. 

That's not enough memory for the use 
of a high level language such as BASIC, 
so the program had to be written in 
assembly language. Since my previous 
assembly language experience was with 
the 8080A, this was the CPU chosen for 
Robin's system. 

We wanted the system to be expandable. 
In the future, Robin may want to add 
more memory, a printer, a speech synthe- 
sizer or other additional peripherals. 
For maximum flexibility in expansion, 
the S-100 bus structure was chosen 
because of the wide range of commer- 
cially available plug-in circuit cards. 
The computer also had to be small and 
light enough to mount under the seat 
of Robin's wheelchair. In order to modify 
the menu and message areas of the video 
display independently, the computer 
needed a memory-mapped video display. 
These constraints pointed us toward the 
Polymorphic Systems' Poly 88 System 4 

The Poly 88 uses a 5 slot S-100 chassis, 
which makes it small and fairly light in 
weight. The Poly video card is memory 
mapped and displays 16 lines of 64 
characters each-just right for Robin. The 
features of the Poly CPU card were also 
useful: it has 512 bytes of RAM together 
with a monitor program in ROM. A 
cassette tape interface card works 
together with tape loading software in the 
monitor ROM to handle program storage 
and loading. 

The vocabulary for Robin's system is 
stored in RAM because we expect her 
vocabulary needs to change once she can 
communicate more freely. The problem 
with storing vocabulary in RAM is that 
RAM is volatile— the memory and thus 
the vocabulary are erased every time the 
computer is unplugged. So a battery 
back-up card was added to the system. 
This card keeps the program and vocabu- 
lary stored in RAM even though the 
computer may be unplugged for hours at 
a time while Robin's wheelchair is moved 
from place to place. Robin's computer 
uses the Seals Electronics BBUC card 
with NiCad batteries. 

We had, at one point, considered battery 
powering the entire system, but ended up 
rejecting the idea. A large and heavy 
battery would have been required for 
reasonable life, and this would bring the 

total weight of the wheelchair and system 
up so high that Robin's mother wouldn't 
be able to lift it in and out of their family 
van for trips to school and other errands. 
As it is now designed, Robin's system has 
to be plugged into a wall outlet to 
operate, but the battery back-up card 
keeps memory alive while the system is 
unplugged so that it is instantly ready 
to start upon being plugged in. 


A few additions and modifications were 
made to adapt the commercially available 
hardware to Robin's application. The 
Poly 88 chassis has only two controls: an 
on/off switch and a reset pushbutton. 
This is because it is designed to use a key- 
board for functions which a control panel 
might perform. The reset pushbutton 
starts the ROM cassette tape loading 
program. I added a second pushbutton 
which activates a vectored interrupt and 
jumps to the beginning of Robin's pro- 
gram. This makes it possible to start up 
Robin's system without the keyboard. 
A schematic for this simple addition 
is shown in Figure 1 . 

Figure 1 

As a computer powers down, it can 
scramble data stored in memory by 
sending out false write commands. To 
eliminate this problem, the memory in 
Robin's system was partitioned so that an 
8K block of RAM, containing the main 
program and stored vocabulary, could 
be write protected. This left only the 
512 bytes of RAM on the CPU card 
unprotected (and the memory mapped 
video display, of course). The small 
CPU RAM area is used for all scratchpad 
functions and is one of the features of 
the Poly CPU card which encouraged 
its selection. 



BOX 1579. PALO ALTO CA 94302 


74 LS 

, , _ 00 



MWR- (PIN 3, ALL 21 L02's) 

100 pF -^r 

Figure 2 

43~ > > DI7 

SINP | 46 
A15 | 32 > Q 

Figure 3 

The RAM card used in Robin's system is 
an Industrial Microsystems IuS #000231 
8K card which uses 21L02-4 chips. This 
card was modified slightly so that a 
toggle switch could be added to the 
computer's front panel which protects/ 
unprotects the main 8K RAM. When 
loading new programs from cassette 
tape, RAM is unprotected. Otherwise it 
is protected. A schematic of this circuit is 
in FiguiC 2. 

The final hardware modification for 
Robin's system was the addition of an 
input port for her kneeswitch. Figure 3 
shows the schematic for this circuit, 
which was built on a small scrap of 
Vectorboard and mounted on the Poly 
88 chassis. 


The program for Robin's system is listed, 
with comments, on the following pages. 
It was kept as brief and simple as possible 
to leave as much space in memory as pos- 
sible for the storage of vocabulary. The 
vocabulary is stored as ASCII, with one 
;haracter per byte of memory. ASCII 
doesn't use the eighth bit of an eight 
bit word, so I used the eighth bit as a 
'beginning of word' flag. The first charac- 
ter of any character, word or phrase 

stored in memory has the eighth bit true, 
and all following characters (if any) have 
the eighth bit zero. This scheme allows 
the words in Robin's vocabulary to be 
packed tightly in memory. The only 
extra bytes of memory used are flags 
inserted at the end of each subgroup 
(FDH), group (FEH) and at the end of 
the vocabulary (FFH). 

The main program uses one subroutine 
from the Poly 4.0 monitor ROM. That 
routine, WH1, outputs a character to the 
video display. It uses a location in the 
CPU board RAM, POS, to store the next 
position it will print into and it recog- 
nizes several control codes: 
0DH = carriage return and line feed 
0CH= erase screen and send cursor home 

(upper left corner of screen) 
0BH=send cursor home without erasing 

18H= erase current line 

The starting address for the memory area 
mapped by the video display is F800H. 
Thus, if the control code 18H is in the A 
register when WH1 is called, it will stuff 
F800H into POS. WH1 saves all registers 
on entry and restores them on exit. 

The monitor ROM on Robin's CPU board 
is a slightly modified version of the 4.0 
monitor: at address 0008H a JMP 2000H 

has been inserted so that vectored inter- 
rupt VI6 jumps to the start of the main 
program. This allows a single pushbutton 
to start Robin's system. 

Robin's software was hand assembled 
because I didn't have an assembler 
program to run on her system. The pro- 
gram listings were typed by hand and 
may contain a few errors. 


The TEXT and EDITOR programs written 
for Robin's system are both very short. 
TEXT was used to enter the messages, 
alphabets and vocabulary into her 
system's memory from the keyboard. 
EDITOR is used to modify her vocabu- 
lary and to add to it after the original 
entry. Here is what they do in detail. 

TEXT is entered with a starting address 
in HL. The TV screen is erased^ and the 
system waits for text to be entered from 
the keyboard. Any unshifted letter is 
printed on the TV screen as a lower case 
letter but is stored in memory (beginning 
at the starting address in HL) as upper 
case ASCII with the eighth bit zero. The 
keyboard for Robin's system is a 
Teletype -like keyboard and does not 
have lower case letters, so the 'unshifted' 
letters are actually upper case, but TEXT 
translates them for display purposes. 

Any letter of the alphabet typed while 
the CTRL key is held down (except Z) is 
printed on the TV screen as a capital 
letter and is entered into memory as 
upper case ASCII with the eighth bit 
turned on. This allows the first letter of 
any word or phrase to be identified. The 
Poly monitor program uses CTRL Z as 
a command to enter its front panel mode, 
so this is the one exception to the rule 
stated above. Shift O jumps to the 
EDITOR program, at the current address. 
Rubout erases the last character entered. 

TEXT is also capable of inserting the 
control codes which identify the end of 
alphabets, subgroups and groups. 
CTRL shift L = insert FBH 
CTRL shift N = insert FDH 
CTRL shift O = insert FEH 

EDITOR is a somewhat longer and more 
complex program which allows the user 
to examine the text stored in the system's 
memory. It also allows modifications of 



BOX 1579, PALO ALTO CA 94302 


that text by insertions and deletions. 
If a deletion is made, all of the rest of the 
text (at addresses greater than the deleted 
address) is moved down one memory 
location to close the gap. If an insertion is 
made, all of the rest of the text is moved 
up one location to make room for the 

EDITOR is entered with a starting 
address in HL. Upon entry it will display 
a 'line' of text, beginning at that address. 
At the left end of the line, the current 
starting address will _ appear, in hex, . 
followed by a space. Then the contents of 
memory are printed, up to and including 
the first 'control code' found. Any letters 
stored in memory with the eighth bit 
high will print on the TV as capitals, 
while those with the eighth bit low will 
print as lower case. The control codes 
will print as special symbols: 
FBH = { FDH = } FEH = ~ FFH = ■ 

The EDITOR recognizes several com- 
mands, as listed below: 
carriage return = display next line 
line feed = display previous line 
space = redisplay the current line, shifted 
one character to the left 
NOTE: insertions made by EDITOR 
will go just in front of the first 
character on the display. The space is 
used to move along the current line so 
that insertions (or deletions) can be 
made in the middle of a line. 

Shift = jump to TEXT with HL equal 

to the starting address of current line. 
CTRL shift L = insert capital L 
CTRL shift M = insert capital M 
CTRL shift O = insert FEH 
CTRL shift N = insert FDH 
rubout = delete first character of current 

any unshifted letter = insert that letter 

with eighth bit low 
CTRL any letter except L, M, or Z = 

insert that letter with eighth bit high. 

The EDITOR, and TEXT programs use 
several more subroutines from the Poly 
4.0 monitor ROM. Either program is 
entered with a starting memory address 
in HL. The monitor program allows 
register pairs to be pre-loaded from the 
keyboard while it is operating in the 
'front panel' mode. For a detailed expla- 
nation of this procedure, see the Poly 
system manual Volume 2 p58-65. The 
other subroutines used are: 
WHO = fetches a character from the key- 
board and returns it in A. No other 
registers are affected. 
DEOUT = print the two byte number 
in DE as a four character hex number. 
MOVE = move — BC bytes from the 
area starting at (HL) to the area 
starting at (DE) — only works for 
moving to lower addresses. 

Robin's main program turned out to be 
shorter than expected. Including the 

alphabets and punctuations symbols, it is 
1250 bytes long. Even with 1,200 words 
of vocabulary in memory, there is still 
room for TEXT and EDITOR to remain 
in memory so that Robin's family can 
revise her vocabulary as needed. 


It may be helpful to briefly mention how 
the initial vocabulary for Robin's system 
was chosen. The first 1,100 words were 
supplied by Robin's tutor, from lists of 
the first words taught in English. The 
remaining words were chosen by Robin 
and her family. These include the names 
of people, places, articles of clothing, 
foods and other objects which Robin 
comes in contact with. 

As practical experience with the system is 
accumulated, revisions may be made in 
the initial vocabulary and possibly in the 
main program. For example, it may turn 
out that Robin will feel more comfort- 
able spelling words than looking them up 
in the stored vocabulary. If this is the 
case, we may try adding a set of look-up 
tables for prefixes, roots and suffixes to 
speed up communication. 


The basic system built for Robin can 
be expanded and modified to fit a wide 
range of possible situations. For example, 
the kneeswitch could easily be replaced 
by an elect«>myog«ph(EMG), an instru- 
ment which measures the electrical signals 
associated with muscle tension. An EMG 
can easily detect levels of muscle tension 
which are too weak to control a switch 
mechanically. This is a practical alterna- 
tive to the kneeswitch for people who are 
only capable of very limited movement, 
such as an eyelid twitch. There are many 
hospitalized patients who experience 
extreme frustration because they are 
conscious but cut off from communica- 
tion. Microcomputer communication 
systems of some kind may eventually 
become standard hospital equipment, 
and could help to make such patients' 
lives much more rich and meaningful. 

There is a wide range of possible options 
for expanding Robin's system. It would 
be easy to add a printer, for example. 
She could assemble a message on the 
TV screen as usual, and then select a 



BOX 1579, PALO ALTO CA 94302 

'print' command which would cause the 
message area of the TV screen to be 
copied on paper. This would allow her 
to write an essay or a letter. 

An S-100 compatible card is available 
from D. C. Hayes Associates (the 
80 -103 A Data Communications Adaptor) 
which would allow her to select and dial 
a telephone number and send messages 
over the telephone to anyone having a 
computer terminal. A number of com- 
puter networks are now being used as 
communication networks, and it is 
reasonable to expect a network for handi- 
capped people to develop in the near 

Computers can be used to generate and 
control sounds. Several companies now 
offer S-100 compatible speech synthesis 
cards. It would be possible for Robin to 
learn to speak out loud, using one of 
these cards. Although the necessity of 
learning a new 'language' of phonemes 
would initially make this a slow commu- 
nication process, the potential exists 
for this to be a very rapid communication 

Several companies now offer S-100 
compatible circuit boards for music 
synthesis. It would be possible to write 
a program which would allow Robin to 
compose music and instruct the computer 
to perform it for her. Computer graphics 
are also possible. With a higher resolution 
video display, it would be possible for her 
to draw pictures with fine detail, and 
with a suitable printer, make 'hard copies' 
of these on paper. 

There are several CMOS microprocessor 
CPU chips available now. Although 
CMOS memory and peripheral chips are 
still somewhat more expensive than TTL 
and NMOS chips used in Robin's system, 
it is already practical to build a micro- 
computer system similar to Robin's 
which would consume much less power. 
Such a system would be more expensive, 
but would be capable of battery oper- 
ation, increasing portability. 

S-100 compatible circuit cards are readily 
available which allow a computer to 
control electrically operated devices in its 
surroundings. It would be easy to expand 
a system like Robin's to allow her to 

turn on and off lights, appliances, etc. A 
system which can communicate can be a 
flexible control system too. 

If you decide to try to build a micro- 
computer communications system for a 
handicapped person, I'd like to hear from 
you. I may be able to help with advice, 
and Robin might benefit from your 
ideas. My mailing address is: 
Tim Scully 

35267-136 CH 

PO Box 1000 
Steilacoom, WA 98388 

NOTE: Thanks are due to the staff of 
McNeil Island Federal Penitentiary, 
whose cooperation made this project 
possible. The staff of Aquarius Elec- 
tronics in Albion, California were also 
very helpful in tracking down parts for 
Robin's system. Robin's family provided 
the essential financial support, and 
Robin, her family and tutors all helped 
by contributing ideas and suggestions. 

McNeil Island December 1977 






I I 







E E E 

£ £ £ £ fe 

O J5 

"£ 3 

CD .Q 

3 CO io Co C „ Co 

a oj?w a »o."5oo.t; 

^ *i 9 *fc 9 J» "ft 9 .® *R .9 .9 i 



x o 

U O 


(3 Kb 

OW O. C « o 






a. _ 




s£ SiSog d5 O 
zco,_eju.z{n 2 u- z _ 

-J co la. 1 l _l i_j_i>_i v > — X — 

o™-io™-j-jo-j-io-i-jo_iSo z .s<™S 

5 w ^-S z «l.S 8 o s ° s o s o J; 


<D CO 

21 2 

X CL T3 
ep co o 

JK CD ** 

= * £ 



£ ES 

» a 

E £ 

c o 



-i a uj 
-i -Jo. 




^*-«-^"«-CMCMCMCMCMCOMe , JC'JC>CO**< , 5^*** 






<2 ID _T C 

3.E £w 

C CD 3 O 
~:S 8 CD 

I -Si 

~ £ 3 > 


Sfcx £ 

T3 Q. ** 

&:> o s 

* — II * 

•£ °I c 

** co UJ CO 

S"Su- » 
So • S 

o 5 •■§ 

2 o . a 
■w O +* 

■a "° *- « 

- D Z X & 

.E uj S £ o 
sS w« tS 

a»_ O ** c 
c O *• c 

— « e « 
$ -2 « 5 c 
5q"° g £ 

£ c ~ o 3 

9 » .O 

« aS « „ 
h <p .2 .£ -B 




co b S co 
o a- o o 

_J — 1 _l — » I— 

X X X X Ul 
CO -1 CO -I X 


CM < CM < g 





to to io co to 
p o o o o 




BOX 1579, PALO ALTO CA 94302 

t- S S 

• 18 


f« ** 

w -•= 

° I- 


*» „ •> 

m — C 
5 -• © 

§• .s 
• € 2 

Si o 

fc «5 ♦- 



** ** 

a S 

*•> ** 

«• •» 

c c 

o o 

o o 



*- cm 






i J 

1 1 



£ e 

T> O 
g JC 

• » 
























O O 

UJ Jg § co to § 



I I 

i_ c = 

II i I I I § 



2 c 
c • 

§ Sea 
I J™ 





»-x»- 3: «-x»-x*- 


-i — _i — -i — _i - _ij- 




« CM 




o o c 

o O ° 
a 5 ■S 

= c 


I 355-° 

: o "o > 

tTJBS 5 

a • " - 


■2 : 

5= a © 


■* * 1 1 " £ 

P O * > o • 

J c g S *€ 

j itilli 

E 3£JN g 

.O 3 

a S 

quiz o 


CD "J ^ 


■ > 



-JJj>J x >-Q<t N(r - 


p 2<og Q < 

? N 5 r S > ? Z 2 







< o$ 2s cj<j Qz M 5 


S m UJ co 

co S — a 





OC •* a -J £ .5 ** 

w Mc.Zz 

Q z 3 8 «** 

5 iu c A ™ E c 








< cab?n 




O CM N " CD a CM rtl U S O 

BOX 1579, PALO ALTO CA 94302 









» II 

Z 9 

LU "8 o) 

flag in 
ontrol c 

co 5; 

S OT3 

— « 

2 a 





tt CD 

« a 










£ 15 


CD • 

- 8 



S: « 



V rf X 5 ..* h: UI uj u. O UJ 




u. IO 

O *~ O 



10 u. 00 u. u) eo co 35 co 00 


« o C O 

£ "5 .£ c 
» « = " 


sO ^ C 

* «Q =1 
«° go 8 

CO Q. > fl) 

M -i 

-2.TJ g 
n _ o> c 

"* £ o -° 

o "o S> » 

§15 I 



s °-8 i 

3 0) c » 

** o e 
« 0) 5 c 

"~ Q. 


o£> 2 
if c « 

3 ® *~ !? 

♦.*S £ 
x " - S 

— a 

* * 

- E « 
s — ■£ 

— •» c 

+* O m k- C 

~ a £ c » 

c £ „_ a *z 


at 03 

C O 

'5 "o 

.i § 







a. a. _ o. ;j>- 

5 co S 2 5 * X X a. 
oOIo< 0-2 5=2 

'jo ' So tSoo ^ 


O ® e o 


O u. 

aco ui 


3 * to 

« (0 <o 

^ *~ ^ 


»- O t- *- O O O 



<- O CO «~ «- CM *M CM qj 





o O 

■o z 

<B -J 

a -J 

a lli 

O j J; 

C w ® 

E It 

3 « TJ 

K » e 

£« S 





x cc 



co uj 
tt a. 
o o 

Uj CO 

S uj 


o _. 

w a8 



















ui s 


1 _i m 




"gJQJS jB< go 


S .5 J_ 


I E 

o £ 

o o 

c 5 

4- a 

*! I 

as +; to 

C ,^ to 

03 o e 
x: □> a> 

♦* c •- 

5 c 2 


H5 J5 c 


_» ^ 

■o 3=5 






ui I 

o o 

- - 1 < o 

. jnSqqj<Q 
>j^ N jjj-th 


i =■ S o o ^ 5 « o 5 o oc 

-)_J < Q Z OQ 

UJ < ^ S 2 X I 



.E 03 fe 

*■ "5 a 
^ui .2 « 

< i_ *- 

"2° a 8 



** ' 


03 : 

> , 

c * S 

•- »- 2 

.2 g 

I- E 03 

- .is 

a <- 

» o a 

0) co •— 

ffi O 

o r uuS H 00 »- o o 

g„ Mg5g ° og noS 

§^§SSSS^ ^ gS S "cog 

n U|OUICM<CMQui<Q) Z ~Ouj<CM<CMC» 

cmSou-OcmcmOooSO c sOcOCOOCMCMO 

^ o 




s o « 






M Ou-«-cMi0com 




£ «| :o 

o ** *: ^ 







JE tt °§ 

♦* — — > 

g «o w » I- 

S c » S X 

s_~-8 S;Ui 

o c a> o Z 

• 3 

-C to - - 


o " 

O «- 

"".x^o S - 


co £ 
— a. 

"g » = s 

g co 


U. ^C0 

5 . Ogui 
O ir .2 O) 
= 5 o^: 

<C0fflUJCM< n ui 

CO < 

ui O 

O- »TJ 





BOX 1579, PALO ALTO CA 94302 

. UJ o 
C < in 

OS 8 * 

£x _ii- 
oX i3 z 

v -J O -s 

M O 

x O 

» o 


_| CM O O 



t CO W i- 
p: Ui UJ u. 
W *- «- »- 



s jJi 

o. a 

3 3 


0) o 

s * 



CO _l LU 
Z -J CO < 

O OC tf «5 "J 
O uj nr ** ui co 

-IUJI-jH-Q-'- , < 

Jo. In-JOCOC- 






Z uj 

z 5 

T- UJ 

■ O 
uj 5 

< CO 

a. uj 
co s 

X -J 


-I o 

J3 r^~ 

2 UI 

X w 

5- X°-UjJ2ui 

S'JS z .s> 


N C -i 0. - -J N 
Z | < S >< < Z 

-» 2 O -j -J O -) 


o a 






w 5 



< GC X 
£ CO Z < X 

"i3"oJo "2 

^ 0. IC0-JQ» — Q. Q - Ct - a -Q- I r- 

^5X< n <2^2"OOOOxxuj 


< JO 

co a. D 
ui co a. 

*- 1- 

X co ui 
I- *co 





rv. 00 

Z f» 
, UJ "" 


Sir J S§g 

> uj oc *- co 
co Z ui *. Q 

N CN {? CM 2 fiCM ll N 

s s s t g S o 

Qcj^Qw<QQuiuiS m m™QcnSq«?Qcn 

oo u ;cncnN^.^u. co 


^ N M N N N (VI £ 

°»-0> u c92uj<o no 




t- UI 





■- c 
.2 5 




CNCNCJC0C0OC0T* , 'S , ^^ , ^-^-**-'<fr^- 




•z "8 


put it in HL 
r new alpha 
t the chosen 
loop back tc 
r alphabet 
i chosen ite 


£ s 







a — 


te 2 *- 

•a g = -o £ « 

■o U 


•S •'"S « 

ro o a as O h= 

£ Si 



o cb a xi 

C "O 

-Q ^ 

in w 

a> .E 

o o 

03 £ 
-ri *-> 



o z z 

OStujUJ 5 ^ rf ,^ ujco to 

coO<coO < -§-§ coo-o. 

^ 1 -J K- _i Q ^ _i " < _l -JQ^Q 



q a 
z z 



O <U 

a> 3 



Sf 5c\«0 

O CJ o o 

.. v „.. w O O NO 

8<guiS * o 1 * uiuJ 





nnco^roUJnco ffi 33roQ<wQr^ujQc»Q<CQ<NCi) 

w ** UL CM LO 

oj ■» oo a> eft 


00 UJ LL O 
0> 0> OJ < 





C «> 

o E 


o 5g 

Z Z 3 

? 3 a 
a. q. q, 

ui ? ■£ 








J3 O 

3 « 

z e 

"j 0) 

e UI °» 

*■ O CO x 

^: T 






£ 8* E 
I Jog 

O ui . ^ 

XU t Q q. 

z » o ^ O . z 



l" J o x" -I -" UJ z 

x< N x<<s2 


r- a 





BOX 1579, PALO ALTO CA 94302 

° 8 

I 21 -J I P 

CD 3 CO Z 00 > 

U. Q U. < U. -I 

co in co J2 co — 

O < Q uj Q m 



QQ Q- 


CQ < 

Q uj 




"- UJ 

CO h- 

a o 














D oo 

u_ co 

z o 












UJ 00 

O co 






~ U- **• 












03 UJ 






_ _ u. O 

O m in m uj 

a i 
















,« 2 ,« < 

CQ CO 00 — 







00 < CO 

Q uj Q 

« o 

Q. O) 


>■ 00 
3 2 
I Q. 
oco 5 CO 

Ol CQ O 00 
uj Q UJ Q 

CO o 







O 00 03 

< CQ < 
111 Q < 



2 g «- 

c c ~ 
o o a 





1 1- I 


— _ ^ ^ Z I fP U. 5 

-I £ => fe Q <£ O g £ Q | 



oQO(Nc y )<ooir)cDu.o<ooto(o c ^Qo<-muLOcou.Qa) 



V _ «- «- cn cn co co 25 in to r» r- 


O u. u. u. u. 
o co co co oo 

CN ft CN S CO ¥■ CN 8™Q 





«o-CO <* 
3 CO CO 



Sl5TilI?I I ln 1 Hl5J? I t 






~ O i- 



N O 

^ roUJ 

a « n 

>r x 


_i o ^ 

SCC c Q. 
a> >- 

co a_ 

„ CQ 
3 U 
O I 

JC — 




2 r- 

<B CO 

■° < 

00 ^ 

CQ < 

a uj 










m uj 

I m 

u. < 
Q uj 





00 Q 


oo CO 

a cc 













m x 
Q uj 

LL O U. p 


CO oo 
u. u. 



inu.o<oocNcocor-.Q = -5"r:JN2><<p i 7 


BOX 1579, PALO ALTO CA 94302 

a. 8 

3 .O 

* a 

(J Q) 


4- a 

c c 

3 = 

5 3 2 
„ 5= o» .2 H 

2; o-E 5 X 

» £ « ™ Q. J-- 
2 g-o|^oo 

^ .* *i -n -5 


jg o 

8 S o 

^- ffl fc 

- S 5 

2 w ° C " 

™ c i s ° s 2 

S; o ** ° i_ *• 

■ o c 8 c 

o"? .5- 2 

•= t: .e ■= 

. T3 •*■ - 

n^ i 5 » it ,2 S: .2 *: 

«> J C ■— (B 2 CO 

g « 8 

■tr 8 ° 

8.E * 
o t 8 

— W .C 

to +* > 

E£ g 


jn co 
3 o> 
A c 

CD •= 



xx>!ts XQ .!f 

o o 5 ft ° z s ft N ft" ft z ft 5 ^ 

3E S2 ± « £ 5 & "-. 



N5ioa;i5:iQs§5§x3>x«|o § ^ 

I -I 



CO «■« 

OS o- ■- 

£ 8 S te 

s « « s 

£ 5 ♦■> 






it upper 
int as is 
it lower 
not, prii 
ake it lo 
d print 




a ai 

a> o) 3 Q 


2 5? 

_ O _ 3 — a. 
ft z ft o 9 2 



":u.ii;(i;u-u-u.u.o £ u. u. u. o 9 o 

»g Su.SoSoSu.goSu-Sog u.S 3 8 8 8o5-?o§ 

u. 3 S in «n ^ 5 cn O »- 00 S O Q OO S u. g O Q Q c? S S ? « S S 


S9y u ; o '- fN{, '><0oocoOocM 

n ■ ■ 

~ T. • 


E - 
* 5 .« 8 

* 5 «> g ..a 
« jc -_ I- E fee 


■ * ft ^ 3 ~ 
.« O^ ■§ 

-8 I£ B Jl 8, 

'S s 

«S CO 

CO >- 

c " 


S £ 

co *D 

c <-• 


8 s 
^ a 


— ■" "O 

a» £ * 


co s: 
* $ 

c co — 

5 j,j 

o ■= 




42 8. E 

£■ N 

CO w 

2 °» 
co •= 


o ° 

» fit 

i 8 

*^ « a 

(!_ O" Ol 

■Q g •*= .E > 

B .. E C o 

£ o 2 co P 

♦-JK ° " £ 

£ y a-*- co 
3 S a> ^ 

O-c ffi « S 

.. r>.M-- 

» £ ~ 

c - ** 
.£ 3 X 

5 £ S^. 

■O co •S « 

•■= ft • 

o. f J- +- 
w .- a — 

'"& .£ € .!2 

"t U. O 

, ^g 8 

* J X o ^~ 
00 o S 00 





it » fey 

W r\ 00 ^r 22 O 

ffi §8§S §§ci§8ci§S§ a ®S E E*88&g Sfcfeoi tS g g 8»g^cS2 

^ co cd co Q do n<uiQnNUjQ<0QOmujOuiOui n u<QujN< m nuiN<ii)<-u2mmcaUJtNtfmOnS<Ur\i 




BOX 1579, PALO ALTO CA 94302 



Horace Enea and John Reykjalin 
Heur i s tic s , Inc . 
900 North San Antonio Road 
Los Altos, CA 94022 


Speech recognition permits control of devices as well as entry 
of data to computers. Using the control aspect of speech 
allows an otherwise immobile person to control a wheelchair or 
turn lights on and off across the room even though restricted 
to a bed or iron lung. 

A pilot project in speech control is described which uses a 
model car linked to the computer by radio. Computer programs 
are included. 



Robert S. Jaquiss, Jr. 

P.O. Box 500 

Beaverton, Oregon 97077 

(503) 644-0161 ext. 5617 


There are increasing opportunities for 
the blind to function effectively in society 
because of the increased use of computers for 
data manipulation and information retrieval. 
This is possible because blind persons can 
receive the same information as their sighted 

The rapidly increasing number of small 
business computers enable blind persons to 
work as programmers and clerks. Also, as 
research assistants, the blind can use data 
bases such as NTIS (National Technical In- 
formation Service), IEEE and Lockheed to 
research articles on topics of interest. 

The major difficulty for a blind person 
is determining the computer's response. This 
difficulty can be overcome by the use of 
equipment such as braille terminals, Optacon, 
closed-circuit television magnifiers, and 
speech output devices. 

This paper describes some of the devices 
that enable the blind to read computer out- 
puts and explores the future role of micro- 
processors in this field. 

Braille Printing Terminals 

Various devices have been designed to 
produce bra i 1 1 e computer output . Although 
a number of equipment prototypes have been 
built and demonstrated, the only commercially 
available braille computer terminals are sold 
by Tri formation Systems, Inc. of Stewart, 
Florida. Triformation makes two basic types 
of devices: one produces braille on a paper 
tape and the other on fanfold paper. (See 
figure A-l . ) 

An impact printer, such as a teletype or 
lineprinter, can be modified to produce 
braille. A pad (thin rubber or elastic) 
placed over the platen, or in front of the 
hammers on a line printer, allows impressions 
of the period to simulate braille. The 
braille produced in this manner is rather 
poor, because the dots are not spaced 
correctly and because the printing mecha- 
nisms are not strong enough to emboss heavy 

A modification of this type is available 
for the IBM 1403 printer. While rather ex- 
pensive, this printer does produce good 
quality braille. 


The Optacon 

The Optacon allows the blind person to read 
ordinary print directly from the screen of a 
bright crt or from a hard copy (see figures A-2 
and A-3). In operation, the Optacon picks up 
the image to be read with a camera and focuses 
it onto an array of photodiodes. The user feels 
the ends of vibrating wires, one for each photo- 
diode, to feel the shape of the characters. 
Uppercase type is the easiest to read, followed 
by lowercase type. The plainer the type, the 
more easily it can be read. While the Optacon 
is not as fast to use as braille, it is more 

Closed-Circuit Television Magnifiers 

For use by partially-sighted individuals, 
the closed-circuit television magnifier is basi- 
cally a camera attached to a high resolution 
black and white television monitor. The camera 
is equipped with a special 20X magnification 
lens. These machines also incorporate a reverse 
image capability. This provides white letters 
on a black background which, in some cases, is 
easier to read. 

Unless the closed-circuit television mag- 
nifier is equipped with a viewing table, reading 
of hard copy may be a problem, and the copy is 
not portable. A plotter can be used to make a 
portable copy with large letters. However, this 
is a slow job, and the paper must be manually 
placed on the machine. 

Speech Output Devices 

Speech output is available in various forms. 
Votrax types require a string of phonemes for 
speech generation, while the Compu-Talker type 
requires more complicated software. In certain 
applications, such as obtaining listings of 
long programs, this type of output is not accept- 
able because it would require the user to mem- 
orize entire programs in order to make 

TSI (Telesensory Systems, Inc.) makes a 
type of speech board which has a canned vocab- 
ulary on it. The canned words will be generated 
when the board is sent a number corresponding 
to the desired word. 

New Research 


BOX 1579, PALO ALTO CA 94302 

There is some work being done with tactile 
displays, the blind person's version of a crt. 
To my knowledge, there are no commercially- 
available devices of this type. Some proto- 
types have been built that use air to drive 
pins up to form a braille line. Others use 
the input port on an Optacon so a computer can 
generate uniform characters on the Optacon dis- 
play. This last method solves one of the major 
problems of the Optacon, which is the necessity 
of holding the camera to the screen with one 
hand, typing with the other, and reading from 
a braille coding sheet. 

Microprocessors in Aids for the Blind 

Microprocessors will be invaluable in aids 
for the blind because they can perform data 
formatting and searching at a reasonable cost. 
Braille books are very large. For example, 
the 1959 edition of the World Book Encyclo- 
pedia requires 43 feet of shelf space. The 
Art of Computer Programming , Volume 1, by 
Knuth if almost two feet long. (See figure 
A-4. ) Because of the large size and high cost 
of braille books, attempts are being made to 
store books onto data cassettes which can then 
be read by a computer and displayed on some 
sort of refresh device. With this approach, it 
will be possible for the blind to have books 
available, at a reasonable price, which can be 
rapidly read and/or scanned through using an 

Microprocessors can also be used to inter- 
face display devices to instruments. This will 
be much easier in the future, because of the 
GPIB interface bus that is being used more and 
more in instrument design. 


Equipment now on the market enables blind 
persons to determine a computer's response and 
thereby receive the same information as their 
sighted co-workers. 

The cost and size of braille books can be 
greatly reduced by storing them on data 
cassettes. Microprocessors will play an im- 
portant role in the devices that enable the 
blind to read the data stored on the cassettes. 



BOX 1579, PALO ALTO CA 94302 

Figure A-2 

An Optacon is being used to read 
from the screen of a Tektronix 
display terminal. 

Figure A -4 

The braille edition of The Art of 
Computer Programming, Vol. 1 
by Knuth. 

Figure A-1 

The author is reading the output 
of a braille printing terminal with 
his right hand. 

Figure A-3 

An Optacon is being used to read 
a hardcopy printout. 



BOX 1579, PALO ALTO CA 94302 


Carter C. Collins, William R. O'Connor 
and Albert B. Alden 

Smith-Kettlewell Institute of Visual Sciences 

and Department of Visual Sciences 

University of the Pacific, 2232 Webster Street 

San Francisco, California 


In the evaluation of a new sensory aid for 
blind mobility comprising a wearable tactile 
vision substitution system we have made a 
number of behavioral measures. The initial 
batch of data was collected and reduced by 
hand which pointed up the necessity for auto- 
mating the procedure in order to process the 
volume of data anticipated in a large scale 
evaluation program. For this we have designed 
and built an 8080 microcomputer based blind 
mobility evaluation system, utilizing an ultra- 
sonic triangulation ranging method, which today 
contintinuously tracks and plots the real time 
course followed by a blind person in our 20 x 
30 foot mobility laboratory space. The detail- 
ed path of the blind walker is recorded in 
graphical form on a monitor screen overlayed 
by the plan of arrangement of obstacles in 
one of many arbitrarily chosen obstacle courses. 

BASIC language software is being developed 
to compute, store and display some of the more 
important mobility parameters, including colli- 
sion avoidance, average walking speed, location 
and duration of stops, total travel time, 
travel efficiency and safety. As an example 
of the kind of output this system will be re- 
quired to produce we relate the findings of 
our initial study. 

We have examined the effects of some 40 con- 
secutive practice trials on the safety and 
efficiency of indoor mobility performance of 
each of two blind subjects in a laboratory 
travel environment. After two hours mobility 
experience with only the tactile imaging device, 
blind subjects walked freely at about one foot 
per second on a 65 foot mobility course through 
a room cluttered with furniture, detecting 
and avoiding over 95% of the obstacles (100% 
in half the trials) . Subjects decreased stop 
and search time from 61 to 14 seconds with 
a 120 second mean total travel time. Travel 
efficiency (percent time spent walking) in- 
creased from 63% to 86% during these trials. 

These tests demonstrate the feasibility of 
the concept that optical information alone 
presented on the skin can contain sufficient 
information to permit the blind to avoid 
obstacles and steer a clear path for successful 
indoor mobility. 


Vision is probably our most important 
mobility sense, permitting us to walk rapidly, 
accurately and confidently wherever we wish to 
go. Since this source of mobility information 
is not available to the blind, there has long 
been the need for an effective sensory aid 
permitting safe and efficient travel by the 
blind pedestrian. The latest electronic 
guidance devices (1,2) have not yet been 
generally accepted by the blind community (3) , 
and today the long cane remains their best 
available mobility aid (4) . 

It has been the specific aim of this present 
preliminary investigation to determine the 
feasibility of utilizing wide field optical 
information impressed onto the skin as a mobil- 
ity aid for the blind. In this study we have 
set up a synthetic mobility environment and 
have made a number of objective measurements 
of the mobility performance of blind subjects 
using a newly developed wide field of view 
sensory aid as their only guidance device. 

This preliminary study alone required over 
150 man hours of data manipulation. In order 
to expedite processing of the voluminous data 
expected in a full scale evaluation program 
it has become necessary to devise an automated 
data collection and reduction system. This 
system was designed to follow a person picking 
his way through our laboratory mobility testing 
area. We required sufficient range to cover 
the 20 foot square mobility course in our 
20 x 30 foot laboratory. Range resolution of 
less than 2 inches was required to detect colli- 
sions with obstacles, and this can clearly de- 
lineate individual footsteps. The sampling 
rate must be fast enough to preserve the con- 
tinuity of the information, even with rapid 
motion, without in any way encumbering the 
motion of the pedestrian being tracked. In 
addition, the system had to store the X-Y 
coordinates of each sample of the pedestrian's 
track for statistical analysis of a trip 
through the room, and provide sufficient extra 
machine capability for storing the outlines 
and positions of furniture and other obstacles 
in the mobility course. 



BOX 1579, PALO ALTO CA 94302 

For training purposes, we wanted to provide 
the possibility of performance feedback to the 
pedestrian, in something approaching real time, 
by the operator of the system; i.e., a meaning- 
ful way to permit the pedestrian to compare 
one passage taken through the room with 
another, based on information collected by the 
system, and made available shortly after com- 
pleting a passage. 

The system which we have designed comprises: 
An ultrasonic locator system consisting 
of an RF-ultrasonic transponder to mark 
the position of the blind pedestrian by 
A triangulation (R. ,R ? ) coordinate to 

Cartesian (X,Y) coordinate conversion 
program with an output graphics display 
map of the pedestrian's path. 
A mobility performance evaluation program 
resulting in archival mass storage of 
path (X,Y and time) coordinate and per- 
formance evaluation data. 
Ultrasonic Locator System . The blind pedes- 
trian locator and tracker consists of a 
"wireless tether" similar to the apparatus 
used by Strelow, Brabyn and Clark (11) to 
follow the progress of blind pedestrians in 
their mobility laboratory at Canterbury 
University in New Zealand. However, our system 
eliminates the three long strings which they 
attached to the subject's head in order to 
compute his location in the laboratory. Such 
strings would become entangled in the overhangs, 
columns and other tall obstacles encountered 
in" our mobility laboratory. We have replaced 
the strings with two invisible RF-ultrasonic 
links, leaving the pedestrian completely free, 
with no strings attached. This ultrasonic 
triangulation technique is diagrammed in 
Figure 1. 

T^e components" of this system are: 

1) a transponder (transceiver) worn by the 
blind pedestrian made up of an RF receiver 
which pulses an ultrasonic transmitter (small 
40kHz loudspeaker) . This RF receiver picks up 
the signal from 

2) an RF transmitter triggered by 

3) control electronics which are controlled 

4) the microcomputer and 

5) two microphones which pick up the ultra- 
sound pulses generated by the ultrasonic trans- 

These microphones feed their amplified 
signals back through the control electronics 
to the microcomputer which generates the 
X-Y coordinates of the pedestrian from the 
ultrasonic time delays as described later. 
The components of this triangulat ion-tracking 
system are shown in Figure 2A, and the system 
is being worn (by O'Connor) in Figure 2B. 

The operation of the ultrasonic system is 
as follows. The microcomputer (Figure 1) 
sends a start pulse to the control electronics. 
This sets two flip-flops, and applies a 4kHz 

audio modulation signal for about 15 ms. to a 
low power 28MHz R.F. transmitter. The outputs 
of the two flip-flops start two counters 
counting the output of a 10kHz clock signal 
generated by the microprocessor. The 4kHz 
modulation of the broadcast signal is detected 
by a tuned filter in the R.F. receiver carried 
by the subject. This detected signal is used 
to turn on the ultrasonic ( 40kHz) transmitter 
worn by the subject. The ultrasonic receivers 
each receive the transmitted 40kHz ultrasound 
delayed by a time proportional to the distance 
from the subject to the receiver. The first 
received 40kHz signal from each receiver 
exceeding a set threshold is used to reset the 
flip-flop associated with that receiver. This 
resetting stops the counter whose count is 
proportional to the measured distance. (For a 
10kHz clock, one count represents a resolution 
of 1.32 inches.) The microprocessor reads the 
counters, resets them and initiates a new 
cycle with a start pulse. 

X-Y Coordinate Computation . The ultrasonic 
triangulation data must be converted to 
X-Y coordinates for plotting and convenient 
data reduction. The triangulation data R and 

R_ are converted to X-Y Cartesian coordinates 

in the following manner. 

To help speed the calculations some alge- 
braic manipulation was done to allow one of the 
two equations to be solved without the use of 
square roots. 

Referring to Figure 1: 

X = L - X* where L is the length of the 
testing area, in this case 
20 feet. 

2 2 2 

(a) yc + Y Z = R * 

2 2 2 

(b) Y = R - X 

(c) (D-Y) 2 + X 2 = R 2 where D = the 

separation between 

the microphones. 

expanding (c) : 

2 2 2 2 
D - 2DY + Y + X = R 

substituting (a) : 

2 2 2 2 2 

D - 2DY + R - X + X = R 

2 2 2 
-2DY = R - D - R 


from (a) 

Y = 

2 2 2 
R l + D - R 2 


/ 2 2 
X =Vr - y 



BOX 1579, PALO ALTO CA 94302 

Even more simplification and speed-up of the 
software was possible due to the fact that the 
combination of resolution (an inch or two) and 
range (from 10 feet out to the 36 foot diagonal 
of the 20 by 30 foot room) allowed the use of a 
single computer word of precision in the input, 
provided that the data was normalized for this. 

In operation the system runs at the afore- 
mentioned tenth second repetition rate and its 
resolution is fixed by the clock applied to 
the counters. This determines the minimal 
change in delay, between the strobe and the 
returned pulse detectable in either channel. 
In one of the clock periods chosen (C 100 ms) 
sound will travel 1.32", which in turn is the 
unit of measure used with the 256 x 256 
graphics output for a useful length in either 
direction of 28.16 feet vs. the mobility 
course dimension of 20 feet on a side. 

The X-Y coordinate computing application 
program is written in 8080 machine code and 
runs on an IMSAI (Figure 3) equipped with the 
following hardware: 

- Extensys 64K dynamic RAM card with 
appropriate disables. 

- IBEX 16K PROM board with the (half used) 
Processor Tech. ALS-8 firmware used in 
development . 

- Chromemco BYTESAVER with the application 
and I/O programs in EPROM. 

- IMSAI PIC-8 with the counter-mux. hard- 
ware built in. 

- Processor Tech. CUTS board for "CUTER" 
based mass storage on cassette. 

- Processor Tech. 3P&S board for I/O. 

- Matrox ALT 256**2 graphics board for 
plotting the output. 

- Matrox ALT 2480 board running the system 

Everything is tied together as illustrated in 
Figure 3. 

The X-Y coordinate computing program is 
written in 8080 machine code and is indepen- 
dent of any higher level language or operating 
system. During operation the program behaves 
as follows. An initialization routine sets up 
the stack pointer, a jump table, and the 
address of both the list and a buffer where 
sampled data is kept temporarily. It then 
programs the priority interrupt /programmable 
counter board to interrupt the system at tenth 
of a second intervals, and halts. 

The interrupt service routine controls the 
program; when called it inputs data from the 
previous cycle, issues a reset to the hardware 
and outputs the R.F. strobe. 

The time during which the hardware is ac- 
quiring new data, is used in this manner to 
process the data that is now loaded into RAM. 
In processing, the data is first checked for 
errors and then normalized. The "Y" coordinate 
is then found using eq. 1, that result is 
stored and used in the finding of the "X" 
coordinate with eq. 2. The pair of coordinates 
is then output on the map display, and stored 


on a list which resides in RAM, filling upwards 
from 0100 H. Finally the lists' pointer is 
incremented, and things halt until the next 
interrupt . 

Control of the ultrasonic system is imple- 
mented through one output and one input port 
on the Chromenco 3P+S (1/0) board. The output 
port uses three bits to operate a custom built 
circuit mounted on the Priority Interrupt/ 
Counter board (PIC-8) containing two eight -bit 
counters, a multiplexer to switch between them, 
and buffers. Bit zero starts the counters at 
strobe time (they are stopped from counting by 
the detected ultrasonic pulse)- Bit one selects 
one of the two counters, and bit two resets 
them after they have been read. Input is via a 
single eight bit wide input port. 

System I/O is through a Wyle Computer Prod- 
ucts CRT and keyboard terminal chosen for 
economy. Input from it required a custom soft- 
ware package and, an input port and a status 
bit on the 3P+S board. 

Output to the CRT comes from a MATROX 24x80 
memory mapped video board (which occupies 4K 
of space because it uses twelve address bits to 
access characters by row and column) chosen for 
compatibility with printer formats. It also 
needed a custom softward package. 

Pending the arrival of our disk, development 
tools were Processor Tech's. ALS-8 firmware 
module (8K) for assembler, editor, and debugger 
as well as their tape based CUTER system and 
board for mass storage. 

The applications package has its own special 
output (Figs. 2 and 3) on another CRT graphics 
monitor; this is driven by a MATROX 256*256 
graphics board, which looks to the system like 
four 1/0 ports. They include; "X" address, 
"Y" address (both 0-255) , intensity (on-off , 
but expandable with decoding and multiple 
boards for colors and grey scale) and screen 
erase. A sample output of a pedestrian's track 
on the graphics monitor is illustrated in Fig. 

There appear to be provisions for synching 
the rasters of the two MATROX boards and sum- 
ming their video to get graphics and alpha- 
numerics in one combination display. 

Taking its input from the X,Y coordinate 
computing program's output, the mobility 
performance evaluation program in BASIC lan- 
guage should prove very flexible. Any changes 
or additional relationships to be calculated 
can be simply added at any time. The mobility 
performance parameters which we have found most 
important to be evaluated at this time include: 
. X,Y coordinates of blind pedestrians' 
instantaneous velocity and acceleration 
of the subject 
. average walking velocity over entire 
location and duration of stops 
. length of individual paths between stops 
. total distance traveled by subject 

49 BOX 1 579, PALO ALTO CA 94302 

. number and direction of turns taken by 


. number of collisions with obstacles 
. number of obstacles approached by subject 
collision avoidance, i.e., percent of 
encountered obstacles with which 
subject avoids colliding 
. total travel time 
travel efficiency 
. the productive walking index or percent 
of total time spent actually walking 
. other factors, as are deemed important, 
can be programmed for computation in 
simple BASIC language. 
Examples of such reduced data are taken from 
our original batch of manually processed infor- 

Sensory Aid . The wearable sensory aid uti- 
lized in these experiments (Figs. 5 and 6) has 
been developed over the past ten years (5,6,7). 
It utilizes a miniature, monolithic, wide angle 
television camera mounted on the frame of a 
pair of glasses which serves as the artificial 
eye of the sensory aid. Images from this 
camera (Fig. : ) are electronically impressed 
point -for-point onto the skin of the abdomen 
by means of a 10-inch square array of 1024 
coaxial stimulus electrodes mounted on a flex- 
ible, elastic supporting garment worn directly 
against the skin of the abdomen (7) . The com- 
plete system weighs five pounds including two 
pounds of rechargeable nickel cadmium batteries 
for eight hours of operation. 

A small and versatile optical system was de- 
signed by one of us (CCC) for blind mobility 
use with the sensory aid. The optics provide 
an adjustable field of view up to 180°, an 
infinite depth of field which permits operation 
with no focusing adjustments required by the 
subject, and a large aperture (f:0.5 for oper- 
ation with available room light. 
— ; Prel iminary trials indicate that a 90-degree 
field of view appears to be about the best 
compromise between sufficient resolution with 
the 32-line system to detect obstacles and the 
very wide peripheral field of view necessary 
for mobility. The optical axis of the lens was 
directed 45 downward to include a field of 
view from just above the horizon down to the 
space directly in front of the subject's feet 
in order to permit him to detect low obstacles 
within a footstep of his path. This field of 
view and lens direction were used throughout 
these tests. 

Mobility Course . The experiments performed 
in this study were carried out in a modular, 
quickly alterable and objectively definable 
synthetic indoor mobility environment contained 
in a 20 by 30 foot room. This mobility course 
was layed out in a Cartesian coordinate grid 
system with one foot square vinyl floor tiles 
(Fig. 4A) . The obstacles consisted of real 
walls, a door frame, pieces of real furniture, 
high and low tables, chairs, a podium, waste- 
baskets, boxes and wall curtains, as well as 
simulated columns and overhanging beams con- 


structed of corrugated cardboard for subject 
safety (Fig. 6). 

Sixteen different courses were layed out on 
graph paper with obstacles located on numbered 
squares corresponding to those of the room. 
This facilitated quick relocation of obstacles 
and permitted a rapid and essentially random 
temporal sequence of different layouts to be 
presented to each subject. Illumination was 
30 to 50 footcandles at floor level. 

The mean total path length of a single mo- 
bility course was 65 feet (20 meters). Of the 
30 total obstacles, it was expected that only 
10 or 20 might be closely encountered by the 
blind subjects as they made slight variations 
in their travel paths. The mean number of 
closely approached obstacles requiring avoid- 
ance was 12 per course. The mean free travel 
path length between obstacles was 5 feet. There 
were 6 turns per route on average. 

The mobility course and evaluation design 
philosophy has borrowed heavily from the pre- 
cepts of Armstrong (8), Kay (1), and, in par- 
ticular, Shingledecker (9), in Emerson Foulke's 

Subjects . Two blind male subjects, age 28 
and 37, were utilized in these experiments. 
Neither subject possessed any degree of func- 
tional vision and each had been blind for over 
25 years. Both subjects were excellent cane 
travelers with years of practice. They each 
had over 200 hours experience with other, fixed 
tactile imaging devices. 

Procedure . Subjects were initially given 
about 15 minutes of pretrial mobility experi- 
ence with the sensory aid. They were given a 
verbal description of the nature of the courses 
and were requested to walk at a normal pace 
avoiding collisions and that the course was 
designed to keep the "shoreline" left. The 
experiment consisted of the subjects walking 
completely through xes pec t ively 39 and 48 con- 
secutive trial courses, each one selected from 
the 16 different courses , such that each course 
was different from the last. Successive trials 
were made about every five minutes . Experimental 
sessions lasted from about 15 minutes to one hr. 

The entire series of tests were recorded on 
videotape and the experimental data were ob- 
tained by a number of replays of the tape. Two 
observers with stopwatches and a counter mea- 
sured the distance and duration of each short 
path leg walked by the subject; number and 
duration of stops; number, location and type of 
collision, and number and size of head move- 
ments . 

The detailed path of the subject was record- 
ed in graphical form overlaying the plan of 
arrangement of obstacles for each course as in 
Figure 4A. A separate plot resulted for each of 
the 87 total runs. Data were correlated with 
these graphical records. 

Results . The travel safety of the blind 
subjects was scored in terms of collision 
avoidance, that is, the percent of the encoun- 
tered obstacles with which they avoided collid- 

BOX 1579, PALO ALTO CA 94302 

ing. The collision avoidance score showed a 
mean value of 91.77% for both subjects combined, 
with an initial value of 84.5% and final value 
of 95.04%; an increase of 12.5%. For subject 
B.G. the mean was 89.87%. The linear regres- 
sion fit of the data for this subject indicates 
an initial score of 86.75%, increasing to a 
final score of 92.99% for a 7.19% increase over 
39 trials. The mean collision avoidance score 
for subject L.S. was 93.31% with an initial 
score of 89.91% and a final score of 96.72%, 
an increase of 7.56% over 48 trials. This per- 
formance is shown in Figure 7 with the linear 
regression fit of the data. 

As suggested by Shingledecker (9) , the 
change in travel time with practice was ana- 
lyzed in terms of three components of travel 
efficiency: walking speed, number of stops, 
and duration of stops along the travel path. 
The mean total time to negotiate the mobility 
course was 120.3 seconds for both subjects. 
Subject B.G. took more time at the outset, 
183.4 seconds vs. 108.9 seconds for subject 
L.S. But with two hours training both subjects 
took the same final time, about 98 seconds 
mean to complete the course. During this 
experiment subject B.G. decreased his total 
travel time 46% with a mean time of 141.1 
seconds (Fig. 8) . The faster walker, subject 
L.S., decreased his total time by 10% with a 
mean of 103.3 seconds total travel time to 
complete the mobility course. 

Walking speed indoors was fairly stable at 
3.8 feet per second mean for both subjects. 
(We have observed indoor walking speed to be 
roughly half that of outdoor speed for both 
blind and sighted persons.) The mean walking 
speed for subject B.G. was ,73fps, showing a 
5.6% increase (Fig. 9). The mean walking 
speed for subject L.S. was .84fps over 48 
trials. We could not measure a change in his 
talking speed. 

The mean number of pauses or stops along 
the travel route to search for a new clear tra- 
vel path was 2.95 for both subjects combined. 
They initially made a mean of 5.11 stops de- 
creasing by 80.3% to a mean of 0.92 stops by 
the end of the experiment. Subject B.G.,with 
a mean number of stops of 3.36, showed a 72.8% 
decrease from an initial value of 5.5 stops 
to a final value of 1.2. The mean number of 
stops for subject L.S. was 2.62, with an 
initial value of 4.8 and a final value of 0.7, 
for a decrease of 85.4% during the course of 
the experiment . The mean number of stops for 
both subjects combined decreased from 5.11 
initially to 0.92 stops at the end of the 
experiment, an 80.3% decrease. 

The mean total stop and search time for 
both subjects combined was 37.5 seconds, 
varying from an initial value of 61.3 seconds 
to a final value of 13.7 seconds; a 78% de- 
crease over the duration of the experiment. 

Subject B.G. with a mean total stopped time of 
55.2 seconds decreased 88.3%; from 98.8 seconds 
initially to 11.6 seconds final value (Fig. 10). 
Subject L.S. exhibited a mean total stopped 
time of 23.2 seconds, decreasing 49.8% from 
30.9 seconds initially to a final value 15.5 
seconds . 

The mean duration of each individual stop 
and search period was 12.93 seconds for both 
subjects combined; initially 12.0 seconds, it 
actually increased to 14.89 seconds final value 
(but subjects averaged only one stop at the end 
of the experiment) . 

The PWI, or Productive Walking Index, intro- 
duced by Armstrong (8) , is a measure of the 
continuity of progress of the subject towards 
his goal . It is the percent total time spent 
actually walking. The mean PWI score for 
subject B.G. was 68.2% with an initial value 
of 48.7% and a final value of 87.7% for an 
increase of 80%, an impressive practice effect 
as shown in Figure 11. The mean PWI score for 
subject L.S. was 80.60% with an initial value 
of 74.12% and a final value of 87.12% for an 
increase of 17.48%. The mean PWI for both 
subjects was 75.04%. Initially 62.7%, with a 
final value of 87.3%; mean PWI for both sub- 
jects increased 39%. 

Discussion . The experimental results indi- 
cate that the optical information provided by 
the new tactile television sensory aid has per- 
mitted blind subjects to safely avoid most 
obstacles and to steer a clear path for suc- 
cessful and increasingly efficient indoor 

Apparently the device immediately provided 
them anticipatory information about the travel 
route as evidenced by their initial 85% colli- 
sion avoidance score. We are encouraged to 
believe that with considerably more practice 
subjects could learn to avoid essentially all 
obstacles, as suggested by their 95% collision 
avoidance score after only two hours of prac- 
tice, and 100% in 10 out of the last 12 trials 
for subject L.S. (Fig. 7). 

The blind subjects quickly learned to in- 
crease their travel efficiency with the sensory 
aid as shown by the 39% increase in their Produc 
tive Walking Index and 31% decrease in travel 
time during the same two hours practice. The 
most significant component contributing to the 
decreased time to complete the mobility course 
was the 80% decrease in the number of stops 
made by the subjects, resulting in a 78% de- 
crease in stop and search time. By the end of 
the experiment subjects were averaging about 
one 15 second stop for the entire 65 foot 
mobility course. The relatively long scanning 
time taken by subjects at each stop may well be 
due to their attempts to recognize details of 
obstacles and escape routes with the low reso- 
lution (3°) due to the 90° wide angle of the 
display (90°/32 lines^3°). 



BOX 1579, PALO ALTO CA 94302 

To meet this problem we now have designed 
and built a 16 to 1 zoom lens with a field of 
view from 10 to 160°. This will permit the 
blind pedestrian to utilize a wide field of 
view for orientation and navigation, but when 
he encounters an unrecognized obstacle he will 
be able to focus his attention by zooming down 
onto the details of the object in order to 
better recognize it. Thus, we believe subjects 
will be able to reduce their stop and search 

Because of the inordinate amount of time 
required to collect and process this type of 
information, a great body of valuable mobility 
performance and rate of learning data would go 
uncollected if an automated collection and 
reduction system were not available. We have 
designed a small, economical, dedicated micro- 
computer system to fill this need. 


We especially acknowledge the professional 
collaboration of Mr. Bruce Smith who wrote the 
applications program which will form the basis 
of another article to be published elsewhere. 
We wish to acknowledge the expert profes- 
sional assistance of Mr. Jim Brodale in the 
preparation of the line drawings and of Ms. 
Helen Sullivan in making the photographs. We 
also express our appreciation to Miss Gail 
Matthews for her arduous efforts in producing 
this camera ready manuscript on short notice. 


1. Kay, L:Conf on Eval of Sensory Aids ,NAS, 1972 

2. Nye, P:Prelim Eval of C-4 Laser Cane, NAS, 

3. Grays tone, P $ McLenan: AFB Res Bui 17, 173, 

4. Hoover, R:in Blindness , P Zahl, Princeton 
Press, 1950 

5. Collins :Proc Nat Sym on Info Display 8, 290, 

6. Collins: IEEE Trans Man -Machine Syst 11, 65, 

7. Collins § Madey: Proc San Diego Biomed Sym 
13, 1974 

8. Armstrong, J: Human Factors in Health Care , 
Pickett and Triggs (Eds) Lexington Books, 
Lexington, MA, 1975 

9. Shingledecker, C: PhD Thesis, U Louisville 
KY, 1976 

10. Gibson, J: The Senses Considered as 
Perceptual Systems , Haughton Mifflin, 
New York, NY, 1966 

ll.Strelow, ER, J.A.Brabyn and G.R.S. Clark: 
Behavioral Research Methods and Instrumenta- 
tion, 1977 

This investigation was supported by 
the Department of Health, Education and 
Welfare, Public Health Service Grant 
Number 1 R01 EY00686 from the National 
Institutes of Health, National Eye 
Institute; Grant Number 501 RR-05566 
from the Division of Research Resources; 
and Grant Number SKF1004 from the 
Smith-Kettleweli Eye Research Foundation. 



BOX 1579, PALO ALTO CA 94302 



Receiver No. 2 

*► X 

of Pedestrian 

Receiver No. 1 

Fig. 1. Schematic layout of the microcomputer 
controlled ultrasonic locator system for 
generating the X-Y coordinates of a blind 
pedestrian from ultrasonic triangulation data. 

Fig. 2A. Components of the ultrasonic 
locator system for blind mobility tracking. 
The RF receiver is the small box in the center 
foreground with its trailing antenna. This 
receiver is generally worn on the belt of the 
blind pedestrian. The ultrasonic transmitter 
is shown in the right foreground as a small 
omnidirectional ultrasonic loudspeaker mounted 
on a lightweight headband worn by the pedes- 
trian being tracked. The RF transmitter is 
in the left foreground and the IMSAI 8080 
microcomputer in the lower background. The 
control terminal on the right and the output 
display graphics monitor is on the left. The 
blind pedestrian's path is traced out in detail 
on this monitor. The sequential coordinates of 
this path are stored in memory and subsequently 
dumped onto a cassette recorder (not shown) 
when the 64K memory is full. 

Fig. 2B. The microcomputer controlled ultra- 
sonic locator system showing the components in 
use. The subject is wearing the lightweight 
ultrasonic transmitter (loudspeaker) on his 
head and he carries the RF receiver. The two 
microphones (on stands at the far edges of the 
picture) pick up the ultrasonic pulses from the 
subject and deliver them to the .uicrocomputer 
which continuously computes the X-Y coordinates 
of the subject and plots them on the graphics 
monitor screen (left background) . 



BOX 1579, PALO ALTO CA 94302 










TT\ I I ;=}£ PIC " 




R/S. 4 \ 








n a 

















Fig. 3. The hardware complement of the micro- 
computer controlled ultrasonic locator system 
(described fully in the text in the X-Y 
coordinate computation section) . 



BOX 1579. PALO ALTO CA 94302 

10' 15' 20 


Pig. 4A. (Top) Plan of one of the 16 mobility 
courses with a typical path walked by a blind 
subject. Note he avoided all obstacles 
(except the small basket on the floor) and 
found a clear path through the maze of 
obstacles with only the tactile mobility aid. 

Fig. 4B. (Bottom) CRT tracing of a similar 
path of subject plotted by ultrasonic tracking 
device and objective microcomputer evaluation 








Fig. 5. Artist*s conception of the 1024-point 
fully portable electrotactile mobility aid. 
The miniature TV camera mounted on a pair of 
glasses permits the object at which the wearer 
points his head to be imaged on the skin of 
his abdomen. The image is converted to a 
pattern of electronic pulses applied to the 
skin by an array of small electrodes in a 
flexible undergarment. 

Fig. 6. One of the 16 furniture arrangements 
of the indoor mobility course used for blind 
mobility tests. Course required subject to 
avoid overhang (in background) , and thread 
between chairs and tables. In the background 
is a cul-de-sac of tables he was required to 



BOX 1579, PALO ALTO CA 94302 

^ 100 


















■♦• 4- 4>++-t-+ 4- -#■ 

4-"*-*-+4-4- +4"*-+ >♦• 







Fig. 7. Collision avoidance mobility data is 
one measure of blind pedestrian safety with a 
sensory aid. Here, performance improves over 
48 trials. 


20 30 40 

Trial Number 


Fig. 8. Total time to complete each course is 
a measure of efficiency of blind mobility. 



BOX 1579. PALO ALTO CA 94302 

;uaoJ9d 'ewii ipjeas 

O (D 
CO .Q 


cm i- 


oas/;aaj paads Buj^ibm 

















•H rH 




~. a 


M 4J 

ft «3 

\ + 





X c 


c w 






C -h 



•H 4J 


\ + 
\ + 



ns m 



* ■£ 



\ + 
♦ \ 

\ + 
+ \ 

\ + 

\ + 
\ + 





•H CU 
4J JC 
O -*J 


\ + 



04 -M 

♦ \ 





+ \ 



. 11. 








01 CO H 






•H X! « 

C* -P £ 

IMd '9LU|1 6UI>||B/V\ luaojad 



BOX 1579, PALO ALTO CA 94302 


William F. Jolitz, 212U Parker Street, Apt. 309, Berkeley, CA 9^704 


The design of a Voice Output Adapter for visually handicapped 
computer programmers is discussed. This device, based on a 
DEC LSI-11 microcomputer and a VOTRAX VS-6 synthesizer will 
generate speech from ordinary typed text. Phonetic translation 
is accomplished by a set of rules, instead of a dictionary. 
Although this device uses a VOTRAX synthesizer, and experimen- 
tation has been limited to English, the device will act inde- 
pendently of language or synthesizer type. All software will 
reside in the main memory; no peripheral memory will be used. 
This results in a compact device. 


Attempts have been made to use speech syn- 
thesis to aid visually handicapped people, but 
most have not encountered great success. Often, 
these attempts were too costly, too complex, too 
limited or otherwise impractical. An attempt L13 
is currently being made to break this "practi- 
cality" barrier by designing an inexpensive de- 
vice which couples a Hewlett-Packard HP 9825A 
desk-top calculator (see A-l) to a VOTRAX VS-6 
voice synthesizer. This voice display will 
allow completely unhindered use of the calcula- 
tor by unsighted operators, and will require no 
extra training to use. Since the unit is based 
on a microprocessor, it will be compact and 
relatively inexpensive — two features which 
lend themselves to mass manufacturing. The de- 
signed- system is not language" or hardware "deperr- 
dent. It is possible to have a multilingual de- 
sign and/or interface to other computers or 
calculators . 

Why Voice ? There are different methods 
that can be used to present computer output to 
the blind. A method that has been widely em- 
ployed is using a Braille terminal to translate 
a line of text into Braille embossings, which is 
a direct method of approaching the problem, but 
one with some limitations. Some of the limita- 
tions of this Braille method are that it re- 
quires copious quantities of consumable paper, 
prints at a slow rate and requires Braille 

An alternative method is to use voice out- 
put from a speech synthesis unit. This method 
has not been frequently used, primarily due to 
the high cost of the hardware. However, with 
the recent revolutions in the microelectronics 
area (which made personal computing a reality), 
this is no longer a problem. 

Voice methods seem to complement Braille 
methods, in that they require no consumables 

(except, of course, electricity;, communicate 
very rapidly, and require no training. Extend- 
ing this thought further, Braille methods leave 
hardcopy, while voice methods are suited for 
interactive use. The differences are similar tc 
those found when choosing either a CRT or a 
printer type terminal, and this simile is accu- 
rate enough to predict where voice or Braille 
(or both) can fit into an application. 

In data entry applications, hardcopy is a 
hindrance, so CRT terminals are used. Con- 
versely, CRT terminals are a hindrance when com- 
posing programs, since often you need hardcopy 
to refer to later. Hardcopy has nice properties 
that are easy to handle , easy to transport , and 
hardcopy is useful in discussing program/data 
text among a group. Many computer systems mix 
both types of devices to take advantage of both. 

Another— advantage - of the voice method is 

that it can be used by sighted users to double 
check data entry. Feedback of this kind has 
been found to be very efficient at detecting 
entry errors in a flight research experimental 
In addition, an easy-to-use speech synthesis 
system (as this basically is) allows for mis- 
cellaneous special purpose applications (paging 
systems, etc.) Voice is a good medium for get- 
ting a short message across from a computer to a 

A Sample Session With The Voice Display 
System . In order to understand this voice 
system, let us look at how it will work in use. 
In the following example (see A-2), the operator 
will type a one-line program into the calculator 
which will display "hello" on the calculator's 
display panel when run. The "dsp" mnemonic is 
a calculator function to display text on the 
display panel. As can be seen in the example, 
the operator presses a key on the calculator, 
then the key's name is spoken by the synthe- 
sizer. This is called echo-feedback, and it's 
purpose is to allow the user to monitor text 



BOX 1579, PALO ALTO CA 94302 

input on a character by character basis. When 
the operator presses the store key, the calcu- 
lator accepts the typed line of text as a pro- 
gram step. Upon observing the calculator accep- 
ting the program step, the voice system reads it 
out word by word. The operator can execute the 
program by pressing the run button, where the 
entered program then writes "hello" on the dis- 
play (which the voice system reads out) and 
stops. To review the program entry, the opera- 
tor can choose to "fetch" it, whereby the pro- 
gram step is read out as it is displayed. 

Basically, the calculator's display is ex- 
pressed with speech instead of print. It is as 
if you had a person reading the display for 
someone who was entering a program but could 
not see the display. 

Prototype Philosophy . The intent of this 
project is to show that an inexpensive device to 
aid blind computer programmers can be made. No 
attempt is being made to engineer a product; all 
that we intend to do is demonstrate the ideas 
for such a product with a working model. All 
equipment is stock, with the lion's share bor- 
rowed from various parts of NASA Ames Research 
Center, Digital Equipment Corporation and the 
Sensory Aids Foundation. All software is gen- 
erated in the high level language, £, £33 for 
ease of programming (in systems programming 
applications, £ is the APL of computer langu- 
ages ) . Any additional reduction in program 
size, through using assembly language, would 
mean a larger cost in programming time and 
frustration (anyway the C compiler used gener- 
ates size optimized code). 


Practicality is a main concern in the 
design of this system. Hardware was chosen with 
availability in mind. Although no attempt was 
made to compact hardware, it was felt that 
overly large hardware would obscure the basic 
concept of practicality. Hardware was chosen 
whenever possible to reduce size. As a result, 
all of the hardware associated with this project 
(less printer) is about the size of a breadbox. 
If one were to custom design the hardware (less 
printer) with no major changes, except for re- 
moving redundant and unnecessary circuitry, one 
would probably have a well-stuffed 8" x 10" 
board. Given future (fourth quarter 1978) tech- 
nology, such a board could probably be built in 
a fourth of the size and at half the cost. 

The hardware that is used for this project 
consists of: 

1. DEC LSI-11 microcomputer (see A-3) 

2. VOTRAX VS-6 Voice Synthesizer 

3. Intercept Interface 
k. Triformations, Inc. BD-3 Braille 

strip printer 

LSI-11. Microcomputer selection was based 

solely on processing speed, physical size, and 
high level language support. Processing speed 
is really a function of what kind of operations 
are most frequently performed. For the speech 
synthesis software, it was empirically discov- 
ered that 16 bit pointer arithmetic operations 
would be the most common (J0% of the time the 
software is searching or indirecting through 
matrices, many larger than 256 bytes long). 
The LSI-11 was chosen because it can handle 
pointer arithmetic more rapidly than other 
available choices (Z-80,8o85,6800) . A high 
level language (C_) was also available that gen- 
erated efficient code. 

VOTRAX Synthesizer . Speech synthesizers 
are constantly improving, as need for clearer 
speech is required. The pace of such change is 
so great that last year's products usually are 
surprisingly outmoded by current products. The 
synthesizer unit used in this project is an 
example of this; although it has moderately good 
performance (intelligibility), in a few months 
it will probably be superseded. However, this 
unit was chosen because it was available, and it 
generates reasonably clear speech. It has a 6k 
phoneme sound-capacity with four possible levels 
of inflection (one should remember that a 
phoneme is somewhat of an abstract concept of 
being a basic sound from which words are made 
up. Also, phonemes vary between languages and 
dialects. When a manufacturer advertises a de- 
vice with a capacity of 6k phonemes, this means 
the device has 6k available sounds which mimic 
some basic sounds in a dialect of a language; 
usually not complete in coverage.) 

Intercept Interface . At the start of this 
project, some limitations were placed on the 
interface hardware. One was that no modifica- 
tions of any kind would be made of the calcula- 
tor. In other words, access to the signals com- 
ing from the calculator's keyboard could not be 
made via "pick offs" on the circuit boards of 
the calculator, but instead must be made in a 
more civilized manner by attaching to some con- 
nector already available on the outside of the 
calculator. It was found that keyboard and dis- 
play data could be obtained by subtle decoding 
of the calculator's I/O bus. The device which 
will accomplish such decoding is known as the 
Intercept Interface, which is presently being 
created by NASA Engineer, Donald Billings. By 
using a Hewlett-Packard built card assembly to 
buffer calculator data bus lines, no direct 
electrical connection will be made to the calcu- 
lator from the microcomputer, so both units will 
be isolated. 

Triformations Braille Printer . Rounding 
out the systems hardware completely, a Braille 
printer from Triformations, Inc. allows limited 
hardcopy use. Although the main emphasis in 
this project is to demonstrate voice methods, 



BOX 1579, PALO ALTO CA 94302 

some hardcopy capability is desirable (good 
engineering practice). The BD-3 printer used 
here, embosses Braille on a strip of paper 
(Braille ticker tape!) Something should be men- 
tioned about Braille: Contrary to popular be- 
lief, not all visually handicapped people read 
Braille. Many use low vision aids, which allow 
them to read with what limited vision they have 
left, by hand scanning a portable TV camera over 
text which is viewed on a TV screen. 


The voice system has a large amount of 
software, most of it is concerned with translat- 
ing text strings into phonetic strings. There 
are five major procedures: 

1. Primative - I/O Monitor (PRIM) 

2. Pronunciation by Rule (RULE) 

3. Statement Symbol Translator (SYMBOL) 
k. Display Parser (PARSE) 

5. Braille Code Converter (CODE) 
The interaction among these procedures is illus- 
trated, (see A-k) . Before discussing this fig- 
ure, the individual procedures should be des- 

PRIM . The Primative I/O Monitor functions 
as a buffer to the I/O devices for the main pro- 
cedures. All device-dependent code is present 
here, so that communication between device and 
system software is via queues. Housekeeping 
functions, like device error detection/recovery, 
are also PRIM's responsibilities. By organizing 
the monitor this way, only one procedure must be 
modified to allow for different hardware. 

RULE . To allow for a flexible vocabulary, 
pronunciation by rule was chosen as the method 

to 'convert text into 'phonetic : text, which is 

more palatable to the voice synthesizer. Proce- 
dure RULE will accomplish this by checking input 
text for rules that might apply and performing 
simple transformations on the text when the 
given rule applies (done in real time). Some of 
these transformations are quite simple, like 
removing the silent "e" from the ends of words, 
while others search for sinister medial vowels 
(like the "e" in "houseboat"). A very readable 
article by Allen CUJ describes this process 
well, including both successes and failures of 
rule sets (a translation example from this paper 
is reprinted, see A-5). Failures usually result 
in comprehensible but unusually pronounced words 
(not unlike the way a young child will pronounce 
a new word). A study of a rule system, similar 
to the one that will be used here, shows that it 
produces intelligible speech on 91% of running 
text C5}. Considering that it is almost impos- 
sible to maintain a dictionary of such scope 
(also considering accessing such data in real 
time), this is very reasonable performance. 
This capability is achieved by approximately 800 
rules. The number of rules is limited by 

computer speed and memory storage; it is con- 
ceivable that with faster computers and larger 
memory, performance could be increased. 

In RULE no special processing is done that 
is particular to English language. Only gener- 
alized rules are used, allowing this system to 
be used with other languages. To change langu- 
ages, all that is required is: 1) a new set of 
rules to express phonetic transcription of the 
language, 2) new symbol translation table, and 
3) new voice synthesizer (only needed if new 
language has a different set of phonemes ; in the 
case of Spanish, a trilled r is needed, and in 
German vowels like u are also required) . It 
would be possible to have a multilingual unit 
where rule sets could be selected (probably by 
means of a memory bank switch), provided the 
voice synthesizer has a large enough selection 
of phonemes . 

SYMBOL. A particular problem with most 
computer languages is that they contain unpro- 
nounceable expressions, like */+>=" etc. 
which must be pronounced by use of a symbol dic- 
tionary. This is the function of SYMBOL, to 
pronounce programming language symbols. SYMBOL 
will be large, due in part to the large number 
of program symbols (approx. 200) the calculator 
has (this includes program mnemonics, like prt 
for print, gto for go to, ell for call, etc.) 
In addition to pronouncing symbols, keyboard 
echo feedback is accomplished by this module (to 
avoid unnecessary duplication of code). 


In order to separate program 

symbols from English text that is input to the 
system, a simple LR parser is used to make the 
distinction. PARSE does not blindly separate 
symbols and English, but instead attempts to 
determine if -the -given- text should --be- eons-idered 
as English text, program symbols, or raw program 
data, or a mixture of both. For example, we can 
have the program symbol "prt" (meaning print), 
or the English text "prt" (unpronounceable), or 
program data "prt" (pea are tee). Only syntax 
can decide (not always successfully though) 
which one of these ways the symbols should be 

CODE . A problem in using Braille printout 
for computer use is that standard Braille does 
not have all the special symbols required by 
most programming languages. The disparity be- 
tween standard Braille and ASCII character sets 
is quite large. The differences are critical, 
as a line of program code that might appear 
(in say BASIC) as: P=Xt3 + 20 * X * Y + 3/Z 
would have the Braille form of P=X3 + 20XY + 3Z; 
which Is totally different in meaning. There 
are a few methods to deal with this difficulty; 
the Braille character set can be expanded, or 
the unrepresented characters can be expressed as 
combinations of existing characters. Expanding 
the Braille character set has the obvious prob- 



BOX 1579, PALO ALTO CA 94302 

lems associated with changing a -widely used 
standard (can you imagine all the trouble that 
might occur if ASCII was extended from 7 to 9 
bits /character?) Playing with such standards 
shouldn't be done lightly, as it might have dis- 
astrous consequences for general purpose use 
(for example, making Braille much more difficult 
to learn or use). 

To avoid these problems, one can use the 
standard code and use combinations of characters 
to represent special characters, like using 
"greater than" to represent ">". This has the 
advantage of incorporating all character sets 
that can be described (i.e. in APL, EJbecomes 
"quad quote", but there is difficulty with an 
arbitrary symbol like §J ! ) A disadvantage is 
that it now takes 8-10 characters to repre- 
sent one symbol, which wouldn't be bad if it 
wasn't for the fact that Braille takes up 
four to six times as much space to print 
as standard text. The compromise that has 
been chosen is to use compressed mnemonics, like 
"cln" for ":". 

With the above adjustment in character 
sets, the Braille software allows listing of the 
calculator's program on the Braille printer. An 
ideal situation would be to list Braille and 
typed versions simultaneously on the same 
paper in adjacent columns. This would allow 
easy discussion of the program between sighted 
and unsighted programmers, as each could locate 
errors or discuss critical sections without 
shuffling around. Unfortunately, this takes 
special hardware which is not available. 

by using a set of rules. Echo feedback of the 
calculator keyboard is also done by the micro- 
computer. In addition to the voice system, a 
Braille strip printer will be used to provide 
hardcopy on demand. The complete system will 
be compact and portable. 

How It All Fits Together . The software 
procedures described interact with each other 
along the lines in the Program Interaction 
Graph (see A-U). PRIM acts as a transparent 
buffer to the other four procedures by perform- 
ing I/O functions to/from a device, from/to a 
queue. PARSE reads the current display line 
from a queue, and then separates the line 
according to item type to either SYMBOL or RULE. 
SYMBOL and RULE translate their text into 
phonemes, which are left in the voice synthe- 
sizer's queue. SYMBOL also examines individual 
keypresses from the keyboard and echoes their 
name with the synthesizer. Finally, CODE trans- 
lates any printing requests into expanded 
Braille format and outputs reformatted text to 
the Braille printer. 


A system has been described here which will 
allow the visually handicapped to use an 
HP 9825A desk-top calculator, by means of voice 
communication from a VOTRAX voice synthesizer. 
An LSI-11 microcomputer will translate program 
and written text into phonetic text for the 
synthesizer. This translation is accomplished 



BOX 1579. PALO ALTO CA 94302 


This work was supported by the 

National Aeronautics and Space Administration, 

Contract A^576B. 

Special recognition should be given to 

Susan Phillips of the Sensory Aids Foundation 

for her help in promoting this project. 


ClJ National Aeronautics and Space Administration, Contract AU576B. 

£2J Elson, Benjamin M., "Inexpensive Avionics Concepts Being Sought," 
Aviation & Space Week , August 1, 1977. 

£33 Ritchie, D. M., C Reference Manual , Bell Laboratories, 

Murray Hill, N. J. 07979 

DO Allen, Jonathan, "Speech Synthesis from Unrestricted Text," 

from a collection of papers in Speech Synthesis , edited by Rabiner. 

C5J Mcllroy, M. Douglas, "Synthetic English Speech by Rule", 
Bell Telephone Laboratories, Murray Hill, N. J. 07979. 

A-l Text Accompanying Photo 

"The HP 9825A Calculator is seen here in use. 
The dark window near the top is a 32 character 
LED-ASCII display. The VOTRAX VS-6 Voice 
Synthesizer is the large box sitting on top of 
the calculator. Voice pitch, speech rate and 
volume are adjustable from the potentiometers 
on the front of the case." 



Operator (Keypresses ! 








Calculator Display 

dsp " 
dsp "h 
dsp "he 
dsp "hel 
dsp "hell 
dsp "hello 
dsp "hello" 
0: dsp "hello" 

dsp "hello" 

Voice Output 











store step zero display 

quote hello quote 

run hello 



execute step zero display 

quote hello quote 

Note: THis HP9825 calculator uses STORE as an end-of-line button for program 

statements and uses EXECUTE for indicating end-of-command. 


A-3 Text Accompanying Photo 

"Here is a Digital Equipment Corp. LSI-11 
microcomputer. The five large chips on the 
facing "board contain the microprocessor and 
its control store. In the background can be 
seen supporting power supplies and console 



BOX 1579, PALO ALTO CA 94302 


70 Devices 



3 YNT H E SIZE ?. 




Main Procedures 


















# KRSMS # 

# KROM # 

Reprinted from Allen zh'j 




# M A I A T+ A ALY\JE # 


# DIS = PA SSION + ATE + LY # 

# BSSSPSS ••• 5 # 

# D*IS=P"A A SYN+YT + LE# 

=weak vowel 
(like "uh") 

" =plosive 

= ,+*=pause 




BOX 1579, PALO ALTO CA 94302 


Susan Halle Phillips 

Vocational Coordinator 

Sensory Aids Foundation 

399 Sherman Avenue, Suite 4 

Palo Alto, California 94306 

(415) 329-0430 


This report describes the deve- 
lopment of prototype equipment to inter- 
face a blind telephone operator to a 
TSPS console. Sensory Aids Foundation 
contracted with Telesensory Systems Inc. 
to build an interface utilizing voice 
output to determine if a blind telephone 
operator could perform the job of TSPS 
operator within the performance require- 
ments set by Pacific Telephone Company 
for sighted operators. 

The system has been successfully 
tested and two persons have been placed 
as operators within the Pacific Telephone 
System. This project will now enable 
blind people to be competitive for job 
placement within the Bell Telephone 

Sensory Aids Foundation, a non-profit 
corporation, and the California Depart- 
ment of Rehabilitation established a 
program to develop new employment oppor- 
tunities for job ready persons who are 
blind. This jointly funded grant provides 
for the expansion of entry level occupa- 
tions through the application of sensory 
aids and the development of new adaptive 
devices. During the first two years of 
the Innovation and Expansion Project, 63 
blind and partially-sighted individuals 
were placed in a variety of employment 

One of the primary engineering pro- 
jects initiated by Sensory Aids Founda- 
tion was the development of prototype 
interface equipment for blind TSPS 
(Traffic Service Position System) opera- 
tors. The goal of this project was to 
open the TSPS operator position with 
Pacific Telephone Company to blind people 
and to enable them to perform the job 
competitively within the standards set 
by the Telephone Company for sighted 
empl oyees . 

The TSPS is the specific console 
used by Pacific Telephone Company long- 
distance telephone operators to perform 
their jobs. Generally, the TSPS computer 

console handles all calls requiring operator 
assistance, with the exception of "informa- 
tion" number requests. Sensory Aids Founda- 
tion contracted with Telesensory Systems Inc., 
Palo Alto, to build two prototypes of TSPS 
console overlays which monitored nixie tubes 
and 72 lighted indicator lamps and buttons 
with voice output systems, utilizing the 
Votrax Voice Synthesizer. The voice output 
prototypes would give to the blind operator 
audible information that a sighted operator 
sees. This information is necessary for 
the operator to recognize the state of the 
console and to service customer requests. 

System Design 

The various elements of the TSPS Inter- 
face System are illustrated in Figure A-l. 
At the top of the figure is the TSPS console 
itself. It consists of the nixie tube dis- 
play and the operator control panel with 
lighted pushbuttons and indicators. The 
overlay is placed directly over the operator 
control panel. Each button on the overlay 
is made from clear acrylic and is provided 
with a photo-transister to sense if the button 
is illuminated. Because the buttons are 
transparent, the status of the console may 
be determined by a sighted operator. This 
is important both in training and when the 
supervisor needs to assist in handling a call. 

The Optical Sensing System is placed 
into the recess formed by the housing of the 
nixie tube display, as illustrated in Figure 
A-2. This system is composed of 12 identical 
modules each sensing a single nixie tube and 
thus avoids the difficulties of mechanical 
scanning. The system is both removable and 
portable; an adjustment is provided to align 
the Optical Sensing System to a particular 

The Interface/Control Electronics 
Module accepts commands from the Operator 
Control Box, processes data from the Optical 
Sensing System and console overlay, and 
generates suitable output to a Votrax Voice 
Synthesizer. The output from the Votrax is 
presented to the operator via a second ear- 
phone. With this control box, the operator 


65 BOX 1 579. PALO ALTO CA 94302 

can manually interrogate various sec- 
tions of the console or read the nixie 
tube display, as illustrated in Figure 
A- 3. The entire system is carried on a 
moveable cart with self-contained power 
fupply, as illustrated in Figure A-4. 
The output of a 12 volt DC battery is 
converted to 115 AC by an inverter and 
provides enough power for at least 8 
hours of continuous operation. A battery 
charger is also provided so that the 
entire system may be recharged when not 
in use or at the end of the shift. 

System Operation 

As a call comes in to the TSPS con- 
sole, appropriate lamps light up which 
signal the kind of call, class, and 
charge. For a blind person to service 
the incoming call, these visual cues 
must be transformed into spoken words. 

A step by step outline of the pro- 
cess is as follows: 

a) An incoming telephone call sti- 
mulates a specific pattern of lamps to 
light on the TSPS panel; 

b) The computer recognizes which 
lamps are lit via photosensors on the 

c) The computer determines which 
words should be spoken; 

d) The computer signals the Votrax 
and speech begins. 

Spoken words from the Votrax give 

tire- -blind operator the cues whic.h__t.he__ 

sighted operator obtains visually. These 
cues are necessary to make the appro- 
priate response to the customer. 

Figure A-5 illustrates the TSPS 
console position for two operators. The 
right side has the interface prototype 
equipment for the blind operator. 

Additional Equipment Modification 

A second important part of the pro- 
ject was to modify the job station equip- 
ment, other than the TSPS console, so 
that it could be efficiently used by the 
blind operator. 

Operators must continually refer to 
handwritten notes in order to remember 
names when making person-to-person or 
• collect calls, rate-and-route informa- 
tion for overseas calls, and other infor- 
mation used in filling out manual billing 
tickets. A suitable device was needed 
for blind telephone operators. Equip- 
ment on site could not be noisy. A quiet, 
sturdy, small brail ler was purchased 

for this purpose from the Royal National 
Institute for the Blind, London, England. 
Information stored on plastic cards used by 
all TSPS operators, including area codes, 
operator codes, numbers of business offices, 
repair services, and emergency numbers were 
brail led in a format designed by the Sensory 
Aids Foundation Staff. A template (Manock 
Comprehensive Designs, Palo Alto) was designed 
to enable the blind operator to complete the 
Mark Sense computer ticket. The blind opera- 
tor was then able to transfer information 
from the braille notes to the billing ticket 
by lifting the template over the ticket and 
using a Mark Sense pencil to mark appropriate 


As a result of cooperative efforts 
between the Sensory Aids Foundation, Tele- 
sensory Systems Inc. and Pacific Telephone 
Company, adaptive equipment and job station 
modification to enable the blind to function 
as TSPS console operators have been success- 
fully developed. Using prototype console 
overlays, two totally blind individuals have 
been competitively placed as TSPS operators 
in the Mountain View, California Pacific 
Telephone System. With the success of this 
pilot project, it is anticipated that the 
TSPS interface equipment will permit the 
blind to be hired as telephone operators 
throughout the Bell System. 


I wish to express appreciation to A.J. 
Sword, M. Linvill , "J'V Azevedo and D. Farr 
for providing technical assistance and photo- 
graphs, and to C. Anderson for secretarial 


Figure A- 

Figure A- 
Figure A- 

Figure A- 
Figure A- 

1 Prototype TSPS Interface 
Equipment - Schematic 

2 Nixie Tube Display Reader 

3 TSPS Console with Overlay, Nixie 
Tube Reader and Control Box 

4 TSPS Interface Equipment 

5 TSPS Console Position 



BOX 1579, PALO ALTO CA 94302 















Figure: A-l 

Prototype TSPS Interface 
Equipment - Schematic 



BOX 1579, PALO ALTO CA 94302 


Nixie Tube Display Reader 


TSPS console with overlay, nixie tube reader & control box 



BOX 1579, PALO ALTO CA 94302 


TSPS Interface equipment 


TSPS Console position 



BOX 1579. PALO ALTO CA 94302 


J.S. Bruqler, Ph.D. 

Vice President Engineering 

Telesensory Systems, Inc. 

3408 Hi 11 view Avenue 

Palo Alto, CA 


Telesensory Systems, Inc. was formed in 
order to provide high technology aids for the 
handicapped, especially the blind. Nearly all 
of our recent projects have utilized micro- 
computer technology. The resultant programmable 
aids provide performance and flexibility impos- 
sible to achieve in the past. This paper gives 
details of five of these devices - the SPEECH+ 
talking calculator for the blind, the TSPS tele- 
phone console interface for a blind operator, 
The Games Center for the blind, the Crib-O-Gram, 
and the LSI Speech Synthesizer. Some of these 
devices will also be demonstrated. 


The SPEECH+ is a hand held, battery op- 
erated calculator developed expressly for the 
blind. Our research concluded that speech is 
the most effective calculator display modality. 
SPEECH+ therefore contains a built-in limited 
vocabulary speech synthesizer which provides the 
blind user with spoken speech verification of 
every keystroke and readout of the display upon 
command. The speech unit also indicates over- 
flow and low battery conditions. In addition to 
English, units speaking German, French, and 
Arabic have been programmed. 

Inside the unit, speech and control data 
are stored on a single 16K bit mask-programmed 
ROM. A custom LSI microcontroller chip accepts 
input commands and looks up the "recipe" for 
speaking the desired word. It then constructs 
the proper speech waveform from the stored 
speech data. To reduce storage costs, a number 
of unique encoding schemes are used. The key- 
board scanning, calculating and speech code 
generation are done by a TMS-1000 single-chip 
microprocessor. The use of three large chips - 
the custom microcontroller, the TMS-1000, and 
the 16K ROM - plus a minimum of support cir- 
cuitry enable a convenient portable unit to be 
made available at reasonable cost. 

Since the speech data is stored in 
ROM, vocabularies are easily changed. In 
addition to foreign vocabularies, other 
speech vocabularies can be generated. To 
try other application besides the calculator, 
a "general purpose" and an ASCII vocabulary 
have been programmed. Potential applications 
being explored include talking elevators, 
talking meters, and talking computer termi- 


The telephone operator's job, being 
auditory, has in the past been very well- 
suited for the blind. The introduction of 
computer-based "TSPS" telephone exchange 
systems having many visual cues for the 
operator has made the job impossible for a 
blind person. TSI undertook a demonstra- 
tion project to develop interface equipment 
which would enable a blind TSPS operator to 
compete successfully with a sighted TSPS 

The TSPS (Traffic Service Position 
System) console used by the operator is a 
special purpose computer terminal containing 
over 80 pushbuttons and a 12-digit numeric 
display. The hardware we developed is 
based on a 6800 microcomputer. In operation, 
special optoelectronic circuits read the 
lighted pushbutton and numeric display 
status information into the computer. The 
computer interprets the input data and 
tells the blind operator, via synthesized 
voice, the information necessary to handle 
an incoming call. Once a call is serviced, 
the computer continues to monitor the 
console to verify that the call was completed 
properly. Since a bewildering variety of 
type of incoming calls are possible, the 
details of the system operation are quite 
complex, and have undergone considerable 
evolution. An important feature is the 
ability of the operator to interrogate 
several important console parameters. 



BOX 1579, PALO ALTO CA 94302 

Two demonstration systems were 
built, and are being used every day by 
two blind operators at Pacific Telephone. 
Quantitative performance measures are 
quite encouraging, and we are seeking 
funding to production engineer the system 
to enable widespread usage. 


The availability of low-cost speech 
technology, keyboards, and microcomputer 
know-how prompted us to breadboard a set 
of electronic games for the blind. The 
breadboard unit met great success, so we 
will soon be producing a limited number 
of games units for use at agencies and 
centers for the blind. 

The unit's electronics consist of an 
8800 processor board, a speech board, and 
an analog board. Of the eight games, 
seven are played on the keyboard. Three 
games are modifications of well known 
pastimes (Blackjack, Craps, Tic-Tac-Toe), 
while four were specially invented ("Skeet- 
Shoot", "Number Run", "Tug-0-War", and 
the "Chain Game"). The eighth game, 
called "Paddleball" is a simulated video 
game involving hitting of a moving "ball" 
heard via stereo earphones. The ball is 
indicated spatially by a tone going up 
and down in pitch and back and forth in 
stereo separation. Scoring in "hits" is 
done with the microcomputer until one of 
two players wins. 


Crib-0-Gram is the only project we 
have undertaken outside the field of aids 
for the blind. The Crib-0-Gram is a 
screening device for potential hearing 
loss in newborn babies. It is designed 
to automatically test infants while in 
the crib, and to flag those with a possible 
hearing problem. Babies that fail the 
test are then thoroughly rescreened at 
six months of age. 

The system is controlled by an 8080 
microcomputer. A sensitive motion trans- 
ducer is placed under the mattress, and a 
small loudspeaker mounted nearby. The 
computer monitors the state of the baby's 
activity, and when conditions are appro- 
priate, turns on a 92db 2-4 kHz white 
noise stimulus. The computer then deter- 
mines through various algorithms whether 
the baby reacted to the sound. In order 
to insure statistical validity, a number 
of tests are given over a 24-hour period, 

some of which are silent control tests. The 
computer keeps score and, at the completion of 
the test sequence, gives a pass or fail indica- 

The Crib-0-Gram System is presently under- 
going evaluation at the Stanford Hospital. The 
hardware is designed, and various software 
improvements are continually being added. 
Workers in the field agree that early detection 
of hearing loss is vital. The Crib-0-Gram, by 
using microcomputer technology, will make 
available an automatic, low cost hearing screening 
technique so that remedial steps can be initiated 
at an early age. 


The human vocal tract can be electrically 
simulated by means of series and parallel 
resonators driven by periodic and noise excita- 
tions. Their parameters (center frequency, 
band width, and gain) are varied as a function 
of time to create the various speech sounds. 
This technique is called "Formant Synthesis". 
In contrast to the techniques described in 
section I, an unlimited vocabulary of utter- 
ances can be generated. Nearly all experi- 
mental and commercial formant synthesizers have 
utilized analog circuitrv. This circuitry is 
prone to tolerance and drift problems and is 
relatively costly and inflexible. 

To circumvent the inherent problems of an 
analog synthesizer, we have simulated and 
breadboarded an all-digital formant synthesizer 
suitable for ultimate fabrication in LSI form. 
This unit is programmable, so can implement a 
variety of high performance synthesizer struc- 
tures. If production volumes can be high 
enough, this synthesizer promises to be suffi- 
ciently inexpensive to find wide application in 
aids for the blind. 


Through the use of silicon technology, 
Telesensory Systems, Inc. has developed a 
number of devices for the handicapped. Appli- 
cations range from serious, such as vocational 
aid and hearing screening, to recreational 
games. Such devices can help erase the some- 
what artificial distinction between the handi- 
capped and the "normal". Our ideas for further 
applications always exceed the available 
resources and time. 


The SPEECH+ voice technology is licensed 
from Professor Forrest Mozer of UC Berkeley. 



BOX 1579, PALO ALTO CA 94302 

The TSPS project was sponsored by the 
Sensory Aids Foundation, Palo Alto, Califor- 
nia. The cooperation and help of Pacific 
Telephone was vital . 

The Crib-O-Gram was conceived by Dr. F. 
Blair Simmons and is licensed from Stanford 

The LSI Speech Synthesis work was 
partially funded by The Seeing Eye, Morris- 
town, NJ. 



BOX 1579, PALO ALTO CA 94302 


Larry Tesler, Xerox Palo Alto Research Center 
3333 Coyote Hill Rd, Palo Alto, Ca 94304 


Some of the more challenging games that can be played 
on a computer require more memory than is available on 
today's personal systems. The paper presents various 
simple encoding technique* that were used by the author 
and a friend to implement an interesting subset of a 
complex game on an 8 kilobyte microcomputer. 

Beyond PONG 

Given the choice of a PONG machine and a pool table, I 
will always choose the latter. Hand-eye coordination is 
somehow more challenging when the whole body is 
involved. Moreover, the feel of the cue and the sounds 
of balls colliding and dropping into pockets appeal to my 
senses more than little plastic levers and electronic beeps. 

On the other hand, when I am home and seeking a bit of 
solitary escape, I no longer turn to Solitaire, crossword 
puzzles, TV, or even (unfortunately) reading. I switch on 
the computer and play a game like Roger Chaffee's 
version of Adventure. 









An hour or two later, after a hundred or more moves, I 
have explored all the twenty-five rooms of the labyrinth, 
found the treasure, lost it, found it again, and escaped 
safely to the surface with my skin intact. 

Hungry for greater challenge, I wish my $800 
Commodore PKT were an $800,000 megaputer, so I could 
play Will Crowther's original Adventure, MIT's Dungeons, 
and other complex games based on the Dungeons and 
Dragons theme. With those versions, I could create 
labyrinths for others to explore, with a hundred rooms, 
multiple treasures, assorted demons, magical objects, and 
elements yet to be conceived. The adventurer could give 
commands in stylistic English phrases ("take jewels", 
"throw rock") instead of by answering multiple-choice 

So I study ROGER'S little BASIC program, marvelling at 
how he fit even thirty flowery descriptions of rooms and 
predicaments into the tiny computer, along with their 
interconnection toplogy, not to mention a selection of 
tasty algorithms for deception, subterfuge, and final 

The choice: buy more memory? a floppy disk? or... learn 
sardine canning... 

A clear case of Aladdin economics. Stuff the genie into 
a bottle. 

more bits per byte 

More bits per byte! No, that won't work. 


The ALTAIR BASIC manual has bunches of techniques 
for reducing program space. These work in other 
BASICs, too, including the PET's. Eliminate REMarks 
(ugh), eliminate blank spaces (retch), use the same 
variable for several purposes (horrors). 

Such travesties against good programming technique can 
only be justified in life-and-death situations. But what 
could be a greater emergency? The goblins must be 

Roger has already squeezed the program quite a lot. I can 
obtain another couple of hundred bytes by making the 
program completely illegible. (Please look the other way, 
this is not for children to see.) 



BOX 1579. PALO ALTO CA 94302 

Encoding the Graph 

Hmmm. The topology of the labyrinth is specified in 
DATA statements. Roger has already saved a lot of space 
by encoding some predicaments ("you can't go in that 
direction", "the giant is here, you'll have to get out") as if 
they were rooms. They are marked specially to indicate 
that after the description is printed the adventurer should 
be forced into a real room, either the one from whence 
he came or some other chosen partly at random or partly 
based on possession of the treasure. 

Each room (and predicament) lists its own number and 
six connection numbers, one for each of North, East, Up, 
Down, West, and South. The connection number is 
either to mean no exit in that direction, or the 
identification of another room. A typical example: 

9020 DATA 2,4,0,23,29,0,0 

There are more than sixty distinct characters available on 
the machine (actually, 128), so why not assign a character 
to each room? Then the connectivity specification is 

9020 DATA BD#X]## 

The D signifies a connection to room D, the ft means no 
exit. In general, A-Z represent 26 of the rooms, and then 
on beyond Zebra for the rest. 

This seems to cut 15 characters of DATA to 7. But 
actually, it is better than that, because these DATA 
statements have to be READ. When a number is read 
into an integer array, it is converted to binary and stored 
as a sixteen bit quantity in two bytes. So the array 
storage for six numbers takes 12 bytes of storage for each 
room, in addition to the 15 in the DATA statement (for 
a total of 27). But when a string is read into an array in 
this BASIC, it takes only 3 additional bytes of storage 
(for a total of 10), because the characters of the string 
are not copied: two of the three bytes are an address into 
the DATA statement itself and the third byte is the string 

It takes about the same size program to deal with either 
representation, so we have made a net gain of 17 bytes 
per room, or 510 bytes for 30 rooms. Enough for some 
dwarves, elves, and magic rings no doubt. 


The connectivity specification is not the bulk of the data 
base describing the labyrinth. Most of it consists of 
those flowery text descriptions: 

9020 DATA 2 , 4 , , 23 . 29 , , , "YOU ' RE IN A 

Though flowery, the vocabulary is rather repetitive. The 
word 'gnome' is used in two different rooms, 'there' is 
mentioned in 7 places, 'to' in 9, 'in' in 13, and the word 
'the' occurs no less than 33 times. Altogether, there are 
181 different words, punctuation marks, and word- 
endings (no, I didn't count them, the computer did). The 
sixty "words" used more than once account for 60% of 
the text. Let's assign the following one-character 
abbreviations to the most common English words and 
endings employed in modern dungeons: 


M=a, « = Detter, L=can, indirection, t = -ed, h=tor, 
G=go, H=here, 1=1, J=giaiu, K=cavern, L=little, 
M=message, N=not, O=of, P=pit, R = rock, S=-s, T=the, 
V=room, W=wide, X = Bilbo, Y=you, a=above, b=but, 
c=climb, d = -d, e=-es, g=get, h=tight, i=in, 
j=guillotine, l=ledge, n=on, o=opening, r=are, s=is, 
t=to, v=chamber, w=was, x=gnome, y=you're, &=and, 
\=through, ?=there, *-=bacK 

We can also make some of the PET graphic characters 
stand for words: ± for 'north'; a high-up line for 'top' 
and a low-down line for 'down'; something that looks 
thick and vertical for 'wall'; and so forth. Over 120 
words can be represented this way in PET BASIC. 
Although some BASlCs only provide 60 to 100 distinct 
characters, one can always encode about 250 words with 
an 8-bit code processed in machine language. Anyway, 
sixty words does seem to be a sufficient vocabulary of 
repeated words. The above room description boils down 
to this: 



Don't tell me -- that's the kind of stuff your teletype 
always comes out with! Good; you already know how to 
read it. After DATA" are the seven characters that we 
discussed earlier for the connectivity description; might 
as well combine them with the string used for the text to 
save a little more space. Next comes the text: y=you're, 
i=in, A=a, and then 6NARROW -- which indicates a six 
character word not in the vocabulary -- then \— =east, and 
so forth. 

Various conventions too detailed for this paper are used 
in both the vocabulary and the descriptions to control the 
insertion of spaces between words and the formatting of 
lines of print. The printing subroutine runs a lot slower 
than it used to, but it is still faster than a person can 

Compressing the room descriptions saves a lot of space; 
but the printing subroutine gets longer and the 
vocabulary has to be stored. The net gain is about 400 
bytes. Compressing other messages printed by the 
program should save another hundred bytes. 

The Bottom Line 

A combination of the above techniques frees over 1000 
bytes of storage. A new room whose description shares a 
lot of the existing vocabulary can be added at a cost of 
only 40 or 50 bytes, so we could add another 20 or more 

I would rather add new twists that are available in bigger 
implementations, such as objects that are found in rooms 
and that can be taken along with the adventurer. No 
more than two objects could be carried at a time. Some 
objects would be valueless, others would be valuable to 
take out of the labyrinth at the end of the game. Still 
other objects would be useful during exploration, each to 
surmount a specific obstacle that can be encountered. To 
determine the properties of the objects, the adventurer 
would have to experiment and take risks. 

Educational Applications 

Although Adventure is cast as a one-player game, it 
should be possible to have several people collaborate. 

74 BOX 1 579, PALO ALTO CA 94302 

One way to collaborate is to sit at the terminal together 
planning moves and developing a model of the labyrinth 
Another way is to explore at different times, all agreeing 
to end up in a particular room after the next session. 
I hose who make it back there may then compare notes 
and try to benefit from the experiences of others. 

With more memory, a multi-discipline educational game 
with a similar structure could be concocted by running 
several simulations in parallel. For example, the 
adventurer is an international trader. His or her business 
partner (a simulated being) is off on a voyage but never 
remembers to write home. The partner must be found 
within a year or the business will be taken over by a 
sinister cartel, or the IRS, or something like that 

The adventurer travels through foreign lands seeking 
clues. He must buy and sell commodities in order to 
finance the trip. The supply and demand of commodities 
is affected by location, time of year, and random factors 
such as weather and luck. As time runs out, the 
opposition erects barriers to progress, and so forth. 

Such a game would be an earth-bound blend of 
Adventure and the game Star Trader (see the book What 
to do After You Hit Return). It would teach some things 
about geography and economics, as well as problem- 
solving, strategy, and planning. 

But all this must wait for cheaper data storage. In the 
meantime, if you'll excuse me, a minotaur is waiting for 
me in the family room. 



BOX 1579. PALO ALTO CA 94302 


Eennis H. Allison, Consultant, Menlo Park, CA 94025 
lee Hoevel, Stanford University, Stanford, Ca £4305 


Few tremes are more common in litera- 
ture than that of the epic adventure. 
In these tales the hero, through the 
exercise of his wit and brawn, over- 
comes all to achieve his objective te 
it treasure, romance, power, or what- 
ever. Even the pulp novel, now an al- 
most extinct beast, owes much of its 
structure and character to the tradi- 
tional epic. 

PDP-11?; Adventure is somewhat smaller. 
Neither is really rricro computer or 
personal computer fare at this point. 

The t 

imi ta 



the u 



1/ el 







of fa 

tors pr 
ter game 

sual "gu 

r diffi 
ance . 
er pap 
ters we 
ities a 
ntasy ga 

al epic 
cvide r 
s. Such 
ting ar. 
ess the 
p" fare 

d & ames 
In this 
er put 

nd probl 
mes . 

and it 
ich para 

d ccm.pel 
number " 
. Seme p 
but the 
of con 

some of 
ems of t 

s modern 
digms for 

are far 
ling than 

or "zap 
rot o- epic 
most ful- 
and in an 

In the 

the pos- 
his class 

The si 
not li 
er eve 
zard s . 
one of 
the ga 
so en 
pi ay in 

rer-like cha 
mi ted to com 
ning I was 
e restaura 
ar a table o 
g just how 
snake, ore c 
And then t. 
my collegue 
cgy graduate 
me one even! 
gaged that h 
al, and spen 
g Adventure. 

rm of Adventure is 
puter folk. The oth- 
dining in a local 
nt and chanced to 
f college women dis- 
one got around the 
f the Adverture haz- 
here is the story cf 
s who introduced his 
student neighbor to 
r.gj the neighbor was 
e went cut , bought a 
t the next two weeKS 


The futurists and pundits of the per- 
-5 on a 1 crtirput tiTtr - wor td- all have a-*— 
rounced that the primary use of com- 
puters in the not tco distant future 
will be recreational. On the other 
haiid, I find nearly all computer and 
video games boring. I can't get my- 
self interested for long periods of 
time . 

The first gam 
long enough t 
Adverture. I 
Crouther (now 
fied ty Ion V 
is written in 
to many mach 
game based 
ideas, but wi 
ments, has b 
son f Pare Ela 
Eave Lebling 
Dungeons and 
of tISF out 
called ME I ( 
are rather . 
about 1P0.000 

e which 
c be be 
t was 

at Xer 
oods at 

ines . 

th sig 
een wri 
ek, Br 
at M 
is writ 
of CO 


got my att 
come ertranc 

written b 
ex PAKC) and 

Stanford AI 
N and has mi 

A much ex 
the same g 
nificant im 
tten by Tim 
uce Daniels 
IT; it is 
ten in a d 
Foth pr 

Eungeons re 

(36 bits) 

ed was 

y w. 

, and 
cal led 
qui res 

on a 

Why is Adventure fun to play. Fecause 
it' is an adventure! The essence cf 
the game is exploration of an unknown 
and clever fantasy world with hazzards 
and treasure. Foth Pdvertnre a r.d 
Dungeon use a cave as a universe; a 
cave in which treasure and adverture 
can be found. Beth games owe much to 
Dungeons and Dragons. 


What makes a game fun? interesting? 
compelling? The answers are most cer- 
tainly hurried in the psyche and the 

.__£^14i-tr-a4i ti en of the .player* .However 

there are certain elements which might 
be considered characteristic. First, 
passive games are net really interest- 
ing. The player must be actively in- 
volved. Second, there must be soue 
fantasy fulfillment; the fo ame must 
fulfill some basic psychological need, 
however obliquely. Third, there must 
be some kind of contest and resolution 
of the associated conflict. Fourth, 
there must be some reasonable mix cf 
discovery and invention. Games and 
puzzles with one solution are of in- 
terest only once. Chess, with its 
complexity, is a game of continual 
discovery and invention. This leads 
to a fifth criterion, complexity. The 
game must be adequately complex thet 
it cannot be known, yet not so complex 
as to appear to be random. There must 
be some higher rules at work, but rot 
all motivations and manifestations cf 
the rules should be clear. Lastly, 
there must be variety? bcredom feeds 
on repitition. That is not to say that 



BOX 1579, PALO ALTO CA 94302 

one should exclude ritual (in the 
sense of folk conventions). These are 
f biidamental fabric of knowledge by 

and the t ame 




the game rraker 
corrrrunicates . 

A game 
ture . 

while no 
be textbook 

follows the 
upon a lorg 

f ji c t r a d 1 — 

proven struc- 
real computer game 
perfect, one can ex- 

pect that some elements cf the tradi- 
tional forrr will be preserved. The 
central figure, the hero the player is 
identified with, is of national, 
international, or galactic importance. 
The setting of the game matches the 
importance of the hero. The hero must 
perform some difficult deeds in his 
quest. Gods, daemons, or other super- 
natural beings rray take an active 
part. The game starts in media res 
with the player di scovering~what is 
happening as the game progresses. 

Everyone car make a catalog of fantasy- 
universes which are potential environ- 
ments for such games. Ore can extract 
the universe of dangerous characters 
and the Oriental Express from the 
classic spy thriller, the starships 
and strange beings of the star explor- 
er, the world and honor code of the 
knights of the round table, and on and 


Scripting rather than programming is 
important* The creation of an epic is 
an ercrrrcusly complex task. And there 
is nc information about to indicate 
that programmers are particularly 
skilled at doing it. The creation of 
a ne\» game should be supported with 
special purpose tools so that it is 
available to non-prcgrarrmers . 

The game author must make a number of 
choices. He must decide upon the game 
universe, establish all its natural 
laws, and create all the uersona 
(players, human or otherwise) who po- 
pulate the universe. He must also 
create these inanimate objects of spe- 
cial significance to the game and dis- 
tribute them throughout the universe. 
It is rather like playing god. 

The real problem here is the r.on- 
detenrinistic nature of the game. All 
persona might not appear in any given 
game? the ending is dependent upon how 
the player responds. It is a bit like 
a TV script for which all middles and 
endin t s are worked out. 

Persona are the most difficult prob- 
lem. Good games need interesting in- 
teresting characters to populate their 
world, both good and evil. ^nt on lv 
this, persona must be able to communi- 
cate with the player in reasonable na- 
tural languge, and the response rust 
be tempered by the character of the 
persona. While many characters are 
rather shallow, others must have some 
psychological depth. 

A most important persona is the alter 
ego cf the player himself. It per- 
forms his commands and observes the 
game universe as his eyes, ears, and 
nose. The alter ego may also have the 
role of conscience, arguing with the 
player when he tries to dc ^something 
out of character. 


Svordplay at a terminal is rot really 
practial. Yet ever the ncn-violeit 
games need to have some way of resolv- 
ing contests. One way is to replace 
the actual contest by an idealized 
one. When the black kright encounters 
the white knight, he need rot actually 
joust; a quick game of tic-tac-toe 
might be equally as fulfilling. Other 
possibilities come to mind: a simple 
trivia quiz, an factual quiz, a rid- 
dle, or a simple number game. 


Perhaps the most exciting possibility 
is what one might call the Electric 
Novel. It is like today's escapist 
literature, except it is a participa- 
tory experience. You, the player, are 
the ^hero. You don't experience the 
hero's decisions vicariously,* you matte 
them. We have not really yet solved 
the problems of sex and violence in 
such literature, but it does make in- 
teresting speculation. 

With the advent of inexpensive voice 
recognition and synthesis units, good 
color graphics, and very very lar t e 
mass storage- devices, the possibili- 
ties are even better. 



BOX 1579. PALO ALTO CA 94302 

Special "Laboratory" Session on Computer Games: 

An Experience in Synectic Synergistic Serendipity 

Ted M. Kahn 

Department of Psychology, University of California, Berkeley 

XEROX Palo Alto Research Center 

This session will be held back-to-back with the session on 
Extraordinary Computer Games. Its purpose is to allow 
small groups of participants to generate and elaborate new 
ideas for various types of computer games in an 
"experience it" environment, without worrying about 
problems and details of specific computer implementations. 
The session will be semi-structured through the use of 
meta-game*, a game which I have devised for generating 
new game ideas through cross-fertilization between 
different fields. In addition, various types of typical game 
pieces and boards will provide stimulus material for game 
structures. The entire creation process will proceed as an 
exercise in group problem-solving, one which will allow 
people to meet and cooperate with each other in an 
enjoyable atmosphere while working on a common task. 

The session will emphasize three important aspects in the 
creation of original games, computer-based or otherwise. 
The process is; — 

- svnectic - Many interesting ideas come as a result of 
making connections between fields which may seem to be 
unrelated (e.g., "Think of a game which involves principles 
of physics and fantasy."); 

- synergistic - "A game is more than just the sum of its 
component parts"; 

- serendipitous - One often discovers unexpected 
results or treasures while originally looking for something 
entirely different. This is probably the most exciting and 
unpredictable part of creative activity. 

This session will be of special interest to teachers and 
parents who are interested in helping children of all ages to 
develop their own ideas, and to game programmers and 
non-programmers alike. 

* meta-game c Copyright 1977 by Ted M. Kahn and 
Moshe D. Caspi 



BOX 1579, PALO ALTO CA 94302 

lersh and Al Ahumada, Aero/Astro Dept., Stanford U., Stanford, CA <)h305 


Because of the similarity between video 
game logic and psychological testing logic, a 
microcomputer designed to facilitate game pro- 
gramming can also be used to present stimuli 
acquire responses, and perform the preliminary 
data reduction required for the testing of per- 
ceptual and cognitive abilities. The capabili- 
ties of a game system—the RCA 1802 COSMAC Video 
Interface Processor — will be discussed and dem- 
onstrated with three tests. The game-playing 
microcomputer is an effective tester when the 
display format requires flexible graphics and 
little alphanumeric information. Low cost 
portability, and ease of programming are its 
principal advantages. Demonstration tests in- 
clude an auditory discrimination test, a Stern- 
berg memory test, and a water jar problem-solv- 
ing test. 

The Testing System 

A minimal system for presenting psycho- 
logical test items must have graphics and audi- 
tory display capability, two or more response 
keys, the ability to measure keypress latencies 
to the nearest hundredth of a second, and it 
must be programmable. From among the many 
microprocessors shown at the first Computer 
Faire, we chose the RCA COSMAC VIP because it 
satisfied these requirements, was cheap at $275 
and was locally available off the shelf. We 
were also attracted to the COSMAC because an in- 
terpretive language was was provided for game 
programming, RAM was expandable to h K bytes on 
board, the I/O port provided a convenient audio 
channel, and its low power consumption and small 
size made it appropriately portable. 

CHIP-8, the interpreter, turned out to be 
surprisingly good for programming tests, because 
it was optimized for games. It takes up a lit- 
tle over 512 bytes of RAM and provides a timer 
and a random number generator as well as display 
control and keypad response collection. The two 
byte instructions look like machine code, but 
are easy to learn and use . For example, 8XY4 
sets variable X to the sum of varibles X and Y . 
The instruction DXYN displays the N byte pattern 
at coordinates specified by variables X and Y. 



CHIP-8 and machine language subroutine calls are 
provided for . 

We have had to write three machine lan- 
guage procedures to have psychological testing 
capability: MOVE, which permits the computation 
of one display while actually displaying an- 
other; DUMP, which outputs a waveform in RAM to 
the D/A converter at a sampling rate of 10 KHz; 
and REACT, which measures reaction time with 
2 msec precision during visual display. 


Psychological tests record a sample of 
behavior in response to a standard test item. 
These items might take the form of questions, 
perceptual displays, or objects to be manipu- 
lated. The behavior is usually reduced to a 
number indicating a response category and/or 
the time elapsed between item presentation and 
response. At the end of the test, summary 
scores are computed. Video games like PONG have 
the same logical structure. The test item is a 
moving dot on the playing field, the response is 
the turn of the paddle knob, and the score is 
one against you if you fail to deflect the dot. 


Psychoacoustic discrimination . Two audi- 
tory waveforms are stored in RAM and are pre- 
sented in random order. The subject's task is 
to indicate which is the target sound (or the 
louder or higher in pitch, etc.). Performance 
is scored by the number of correct responses. 

Memor v scanning . A set of from one to 
six digits is displayed for a few seconds. The 
subject decides whether or not a subsequent 
probe digit is a member of the set and responds 
as rapidly as possible. In this situation, the 
reaction time is a linear function of the number 
of items in the remembered set. The slope of 
this function is a measure of the access time of 
short-term memory. 

Water Jar problem . The capacity and cur- 
rent contents of one full jar and two empty jars 
are displayed. The subject's task is to pour 
half the contents of the first jar into the sec- 
ond. The total number of pours measures prob- 
lem solving ability. 

BOX 1579, PALO ALTO CA 94302 


The water jar program was written by Dr. 
Richard Marken, Augsburg College, Minneapolis, 

MN 5540A. 



BOX 1579. PALO ALTO CA 94302 


Beverly J. Jones, Ph. D. 
Assistant Professor, School of Architecture and Allied Arts 
University of Oregon 

This slide lecture presents a historical review of computer 
art and art related applications in computer graphics from 
191*5 to the present. Most of the images shown were generated 
by large- computer systems and involved extensive programming. 
Those aspects which seem most promising for development by 
individuals or community centers with small computers are in- 
dicated. The research, educational, recreational and economic 
possibilities inherent in using computer systems for art- 
related tasks are briefly discussed relative to the images 
shown. Because the conference presentation depends so heavily 
on slides, the paper presented here represents only a summary 
of ideas. It also includes a resource list of bibliographic 


Electronic technology, particularly the inform- 
ation processing devices known as computers 
have the potential of affecting many areas 
of our lives. One of these areas is the arts. 
Because the computer can generate and man- 
ipulate visual images, I believe it offers 
promise for use by artists, art educators, 
art historians, aestheticians, and museolo- 

The pursuit of this line of thought led to a 
collection of slides depicting the images and 
objects which have been generated with the aid 
of computers from 19^5 to the present. A re- 
view, analysis, and projections based on this 
collection comprise the main body of this pre- 

Computer science is a unique discipline in 
that it has the potential for application in 
many fields in which the computer scientist 
is not necessarily knowlegable. Conversely 
the individual knowlegable in a particular 
subject matter area is not necessarily aware 
of potential computer applications. 1^ 
believe these two statements are especially 
true in relation to the arts as a subject 
matter area. Part of the reason for this 
may be the anti-technological stance which 
has beei> fashionable in the arts in recent 
years. As individuals in the arts become aware 
that human values and considered choices 
based on these values can direct computers, 
that computers may be used to individualize 
images, objects and events; and that com- 
puters need not be used in the mode of mechan- 


ical technology, perhaps a greater willing- 
ness to utilize the potential of computers for 
art applications may develop. (20, 21) 

Computer Graphics: A Historical Review 

In the 19U0's analogue computers were used 
to generate the earliest computer graphics 
which were displayed on cathode ray oscil- 
loscopes. Ben F. Lapofsky and Herbert W. 
Franke were among the pioneers in creating 
these images. (h) An early version of a 
plotting device was the Henry Drawing Compu- 
ter, a modified analogue computer designed 
by D.P. Henry. It produced drawings by 
combinations of pen movements and table move- 
ments. (15) 

Later digital computers were used to generate 
computer graphics using line printers, plot- 
ters and cathode ray tubes as the most common 
output devices. Systems combining analogue 
and digital components were also used to 
produce graphics. Most of these images were 
produced by engineers and technicians for 
practical purposes. For example, William 
Fettner's program created the image of a man 
with 7 movable components using data repre- 
senting the 50th percentile pilot of the U. 
S. Airforce.(U) However, some digital 
images were produced for purely aesthetic 
purposes, such as "Stained Glass Windows", 
a graphic designed by the Army Ballistics 
Research Lab. (15) 

Some of the most effective of the graphics, 


BOX 1579, PALO ALTO CA 94302 

with purely aesthetic intent, were created 
by the Computer Technique Group of Japan. 
In any report of computer graphics the work 
of this group is certain to be included. 
Their transformations of photographs of 
President Kennedy and their interpolations 
such as "Running Cola Becomes Africa" may 
be considered classic examples of computer 
art from this period. All of the indi- 
viduals comprising this group were engineers 
and programmers. It contained no profession- 
al artists. At that time very few people 
with extensive art training were working to 
create computer generated images. One of 
these was Charles Csuri and another was 
Robert Mallary. Well known examples of 
their work include Csuri' s film "Hummingbirds", 
his drawing "Sine Curve Man", as well as 
Mallery's machine tooled sculptures created 
with the program TRAN2.(2+, 9, 11, 15) 

Some of the techniques used to create 
computer art introduced a characteristic 
look to this medium. For example, geometric 
graphics generated using equations to de- 
scribe the form have been used extensively. 
Some of these closely emulate the Op Art which 
was popular during the 1960's. Many of 
these graphics make good use of the com- 
puter's ability to do exact and repetitious 
tasks more easily than humans. Another 
technique resulting in a different type of 
form, was the use of stochasticism or random- 
ization in a portion of the program. Similarly 
some environmental variable such as movement 
or sound was recorded electronically and 
included in the image determining data 
within a program. These two techniques 
illustrate the computer capability to generate 
many forms using one program which includes 

variables . ..which., may be. altered at random - 

or with a preconceived pattern in mind. (2, 
K 1^ 15) 

Another type of form was introduced with 
the technology which permitted digitizing of 
the scanned image of a photograph or object 
in terms of a value scale. This technique 
alone and combined with interpolation led to 
the production of a variety of images char- 
acteristic of computer art.(l+, 8) The 
use of interpolation between drawn images was 
also common. Usually these drawn images were 
introduced to the computer using a light 
pen as the input device. (k t 15) 

In 1968 the first major international 
exhibition of computer art was held in 
London. It was called Cybernetic Serendipity. 
Following this exhibit, more artists began 
to take an interest in the computer as a 
creative medium. (15) Currently computer art 
has taken on an international flavor with 
work going on simultaneously in many countries 
of the world. Artists such as Barbadillo, 

Sykora, Giorgioni, Bonacic, Leavitt, Bangert 
and others are now using the computer as a de- 
signing or executing device in their work.(l, 2, 
9, 16) However, some recent technical develop- 
ments have not been widely incorporated in the 
work of artists and remain evident mainly in the 
province of graphics created by technicians 
for experimental and practical purposes. Ex- 
amples of these are three dimensional shaded 
color graphics and computer generated holo- 
grams. (3, 13) 

Art Applications and Projections 

The attitudes and working approaches of con- 
temporary artists using the computer to assist 
them in designing and/or executing their work 
vary considerably. An examination of recent 
issues of Leonardo or of Ruth Leavitt 's recent 
book, Artist and Computer , will reveal the 
variety of conceptual modes and technical 
methodologies which computer artists are using 
in their work. 

Their projected applications are even more 
revealing. For example, Tony Longson pro- 
poses to use the computer to help him create 
forms to better understand visual perception 
and the creative process. Edward Ihnatowicz 
wishes to create responsive Kinetic sculptures 
in an exploration of the field of artificial 
intelligence. He proposes to use these to 
understand cognition through studying the behav- 
ior of these artificial systems which would be 
capable of simulating natural behavior. Charles 
Csuri suggests that artists could manipulate 
visual displays of statistical data to express 
artistic view of reality relating to social 
problems. In a more conventional vein Patsy 
Scala talks of creating visual poetry with 
-computer generated- videtr images and Herbert 
W. Frank expresses interest in creating graphic 
music. Still other artists continue in an 
even more conventional mode. They retain the 
traditional mode of creating art works while 
using the computer to assist them in design. (9) 

Some aestheticians are using the computer to 
generate images for use in testing the 
aesthetic response. Some are using computers 
to analyze statistical data gathered from 
subject's responses to conventional art works. 
Computer simulations of the style of non- 
computer art such as that of Mondrian, Klee, 
and Hartung suggest experiments to determine 
what factors are most relevant in determining 
certain types of responses to works of art.(^, 

Museologists and art historians began to tap the 
potential of the computer for information re- 
trieval and analysis to aid them in studying 
classifying and caring for museum collections. 
The programs created by the Museum Information 
Network have been \jsed by museums for these 



BOX 1579, PALO ALTO CA 94302 

purposes . ( 7 , 19 ) 

A few art educators are interested in using 
computers for instructional and research 
purposes. Guy Hubbard has attempted to 
use computers in programmed instruction. 
Thomas Linehan has used computers to help 
students understand their own preference 
styles. I have suggested research appli- 
cations for three computer capabilities : 
graphics, statistics and information re- / 
trieval.(6, 10) / 

While a few individuals within the inter- 
national art community are "beginning to/ sense 
the potential the computer has for trans- 
forming the conventional views about art, I 
do not believe anyone has projected spme of 
the possible effects this could have/ on 
society. For example, what new choices 
are available to people for use in education, 
recreation or economic use because of 
art-related computer applications. Whose 
responsibility is it to cultivate an aware- 
ness of these choices and share it with 
others? Exploring one illustration may 
illuminate the nature of these choices. 

Moles in his essay "Art, Cybernetics and 
the Supermarket" noted the potential of 
introducing a variable into the/ computer 
program which results in the magnetic tape 
which runs machine tools for ^ndustry. 
By these means every item to/ come from the 
assembly line could vary slightly, thus 
giving the customer more choice. The 
variability would probablj/ be cosmetic in 
nature, not essentially altering the 
product purpose or functional form. Thus 
the choice would be regarded as an ex- 
ample of marginal differentiation. (lU) 

artifacts would not be bound by the same 
considerations. With the dropping cost of 
/small computers and their growing versa- 
tility it appears that many homes will have 
/ several single purpose microprocessors for 
games or for the control of appliances. This 
approach to computer application is in the 
style of mass production and allows the con- 
sumer little control over tlie product ex- 
cept by veto of non-purchase. A small 
computer which allowed the customer to pro- 
gram many essential aspects of the design 
of his environment would be more in keeping 
with the idea that human choice is important 
in directing the use of computers. 

As Duane Palyka notes, "...the versatility 
of this medium is its ability to handle 
quite varying devices for input or output. 
All that is required is that each device have 
a wire or two containing electrical current 
which varies within a certain prescribed 
range. The rest is within the imagination 
of the individual designing the device. "(9) 
To date drawings, paintings, prints, weavings, 
and sculptures, have been created using 
specialized output devices. Responsive 
kinetic sculptures and responsive environments 
have also been created existing as output 
devices. That the public at large could 
manipulate canned programs and create their 
own programs to operate such devices does not 
seem to me to be an unreasonable possibility. 
In this way computers could allow them to 
regain control of the artifacts and environ- 
ments which surround them daily. This seems 
to me an exciting prospect which has impli- 
cations for education, recreation, and our 
economic system. 

However systems such'' as that used by 
Mallary for producing sculptures, Lourie 
for producing weavings, or the Synthavision 
System discussed by Elin could be used by 
individuals wishing to design and create 
artifacts such as furniture, fabrics, 
and prints which would be unique and suitable 
to their special requirements. If canned 
programs with 'many optional branches were 
used to assist individuals in utilizing this 
type of system very little computing knowledge 
would be necessary. Because machine tools 
operating from magnetic tape output could 
produce the artifacts, little knowledge of 
craft processes would be necessary. (7, 9, 

As computers are presently used in auto- 
mation they serve mechanical technology 
which demands exact repetition for mass 
production and is best served by heavy 
centralization of industry. Small systems 
for individually designed and produced 



BOX 1579, PALO ALTO CA 94302 


1. Bangert, Collette S. & Bangert, 

Charles J., "Experiences in 
Making Drawings by Computer and by- 
Hand" Leonardo, vol. 7, p. 289- 
296. 197U. 

2. Bonacic, Vladimeer, "Kinetic Art: 

Application of Abstract Algebra 
to Objects with Computer-Controlled 
Flashing Lights and Sound Combina- 
tions" Leonardo, vol.7, p. 193- 
200. I97lt. 

3. Csuri, Charles, "Computer Graphics and 

Art " Proceedings of the IEEE , 
vol. 62, no. k 9 p. 503-515. 197^. 

k. Franke, H.W., Computer Graphics , 

Computer Art . Phaidon, New York. 

5. Gips, James and Stiny, George, "An 

TnvtiEL.i £ril J 01'i . f Al gr.v I tuaic 
Aesthetics" Leonardo, vol. 8, 
p. 213-220. 1975. 

6. Jones, Beverly, Computer Applications 

in Art Education Research unpub- 
lished dissertation. 1976. 

7. Kranz, Stewart, Science and Technology 

in the Arts Van No strand Reinhold 
Co. , New York. 197 1 *. 

8." KhowTtoh, K. & Harnori, L. , "Com- 

puter-produced Grey Scales" 
Computer Graphics and Image Pro- 
cessing , vol. 1, p. 1-20. 1972. 

9. Leavitt, Ruth (ed. ) Artist and Com- 
puter , Creative Computing Press, 
Morristown, New Jersey. 1976. 

10. Linehan, T. , "A Computer Graphics 

System for Visual Preference Detec- 
tion and Analysis" p. 90-121, 
unpublished document. 

Pictures with an. Ink Flotter," Computer 
Graphics and Image Processing , vol. k, 
p. 200-208. 1975. 

lU. Reichardt, Jasia (ed.), Cybernetics, Art , 
and Ideas , New York Graphics Society, 
New York. 1971. 

15. Reichardt, Jasia (ed.), Cybernetic 

Serendipity , New York-Washington. 1969. 

16. Sykora, Zdenek and Blasek, Jaroslav, 

"Computer- Aided Multi Element Geometrical 
Abstract Paintings," Leonardo , vol. 13, 
p. 1+09-1*13. 1970. 

17. Thompson, Michael, "Computer Art: A 

Visual Model for the Modular Pictures of 
Manuel Barbadillo" Leonardo , vol. 5, 
p. 219-226. 1972. 

18. Tuchman, Maurice, Art and Technology , 

Viking Press. 1971. 

19. Vance, David, "Organization Profile: 

Museum Computer Network, Inc.," Informa- 
tion , p. 157-159, May/ June 1975- 

20. Weizenbaum, Joseph, "On the Impact of the 

Computer on Society, Science , p. 609-6lU, 
May 12, 1972. 

21. Weizenbaum, Joseph, Computer Power and 

Human Reason , W. H. Freeman & Company, 
San Francisco. 1976. 

11. Mallary, R. , "Computer Sculpture" 

Art Forum , p. 29. 197 1 *. 

12. Molnar, Vera, "Toward Aesthetic 

Guidelines for Painting with the 
Aid of a Computer," Leonardo , 
vol. 8, p. 185-189. 1975- 

13. Phillips, J.W., Ransom, P.L. , 

Singleton, R.M. , "On the Construc- 
tion of Holograms and Halftone 



BOX 1579, PALO ALTO CA 94302 


Ceasar Castro 
295 Surrey Place 
Bonita, CA 92002 

Allen Heaberlin 
5737 Avenida Sanchez 
San Dieao, CA 92124 


It does not necessary follow that high 
quality music synthesizes requires complex hard- 
ware. This paper discusses the hardware and 
software design of a synthesizer which utilizes 
a standard microprocessor and a relatively 
simple synthesizer card. Basically the synthe- 
sizer hardware is used for the high speed data 
processing and the software is used for the 
slower data manipulations. The hardware allows 
the microprocessor to control the frequency and 
amplitude of up to 32 tonal channels. Ampli- 
tude control provides the means of producing 
attack and decay envelopes and frequency con- 
trol provides the means of producing frequency 
modulation of the output waveform. The syn- 
thesizer card has storage space for 16 unique 
tonal waveforms. These waveforms can be used 
to emulate different sounds. Their selection 
is controlled by the processor. The under- 
lining design philosophy of the synthesizer 
was to tax the software as much as possible and 
also to give the software as much control as 
possible. This simplifies the hardware and 
gives the greatest degree of flexibility. 


There are several different approaches 
which could be used to design a synthesizer. 
Each has its advantages and disadvantages. 
As to which is the best depends on the pre- 
defined design goals. Before discussing the 
synthesizer design several different synthe- 
sizer methods will be discussed. These alter- 
natives include both classical analog designs 
and several, new digital techniques. 

Analog . This design, usually consisting 
of oscillators, filters, an other special wave- 
form circuits, has traditionally been used in 
synthesizer and electronic organ design. 
These circuits tend to be simple. Usually 
each circuit performs only one function and no 
wide data paths are needed as in some digital 
circuits. A big disadvantage in this approach 
is that analog circuits, such as oscillators 
or filters, can not be time shared. Thus to 
implement a multitone synthesizer many similar 
circuits must be fabricated. In addition 
it is difficult to implement programmable 

analog circuits. For instance, making an ana- 
log oscillator programmable adds significant 
complexity. If the circuit is not programm- 
able flexibility and adaptability is lost. It 
would be difficult to incorporate new wave- 
forms. Furthermore precise control of fre- 
quency and waveforms is difficult. As the pre- 
cision increases the design becomes demanding. 
Oscillator and filter design becomes critical 
and components must be carefully chosen. In 
addition analog circuits have dynamic range 
limitations which are difficult to improve. 
Digital design can, at least in theory, im- 
prove the dynamic range by increasing the 
word length. To summarize analog circuits 
tend to be simple but have serious defici- 
encies when sophisticated performance is re- 

Commo n Divider . - 1 This approach is com- 
monly used in electronic organs and uses a 
common master oscillator, usually around 2 MHz, 
and generates tonal frequencies by digital 
dividers. Usually these frequencies are then 
passed through different analog filters gene- 
rating various waveforms. There are several 
companies which produce the divider chips, a 
significant advantage of this method. Since 
only frequencies which are divisible into 
the master oscillator frequency can be pro- 
duced, resolution is restricted. Thus 
slightly different frequencies can not be 
produced. This approach is easy to imple- 
ment since divider chips are available but 
the approach is somewhat restrictive. 

Digital Harmonic Synthesis . 1 - J In this 
design each harmonic is separately gene- 
rated with the amplitude and frequency spe- 
cified by a piecewise linear function. The 
implementation has good potential in dupli- 
cating waveforms and should be able to gene- 
rate almost any desired waveform. This 
approach requires a amplitude controllable 
sinewave generator for each tonal harmonic. 
This is a significant hardware requirement 
as there easily may be 10 to 15 required 
harmonics. In addition since the amplitude 
and frequency for each harmonic must be pro- 
vided a significant amount of control must 



BOX 1579, PALO ALTO CA 94302 

also be provided, probably beyond the capa- 
bility of common 8 bit microprocessors 
(6800, 8080, Z-80). Thus this approach 
offers high performance but unfortunately 
high hardware and control requirements. 

FM Generation 


In the FM generator 

approach a sinewave (or possible other func- 
tion) is used to modulate the phase of an 
oscillator. The phase is then converted to 
a sinewave, usually by table look-up. If 
the ratio between the modulating frequency 
and the carrier frequency is an integer a 
harmonic spectra is generated. By changing 
the ratio between the frequencies and the 
"modulation index" different spectrums can 
be generated. This approach is fundamen- 
tally simple as few calculations are re- 
quired and harmonic rich waveforms can easily 
be generated. One objection to this method 
is a somewhat subjective one. Since the 
spectrum is not controlled directly but 
rather through a Bessel function it is diffi- 
cult to relate the arguments to the gene- 
rated spectrums. Thus it may be difficult 
to implement a specific tonal waveform if 
the parameters, modulation index, etc., 
haven't been obtained. In addition if the 
waveform produced is not bandlimited to 
the sampling rate aliasing will occur. This 
will introduce distortion as spurious fre- 
quencies will be generated. The only practical 
way to correct this problem may be to reduce 
the spectrum generated by insuring the spec- 
trum does meet the Nyquist criteria. Another, 
even less attractive option, is to increase 
the sampling rate. Thus the approach does 
have potential but does have some pitfalls. 

._^.„. ...„ .„._ . . „„ ^ Synt .h es -■ s (phase Accumu- 
lationTT This approach has been discussed in 
other papers [4] and has been used by the au- 
thors.^ J This approach is similar to John 
Snell'sLoJ approach except the memory stores 
the tonal waveforms rather than just a sine- 
wave. In this design (see figure 1) a tonal 
frequency phase is generated recursively by 
using an accumulator. A digital word, corres- 
ponding to the phase shift between cycles is 
continuously added in the accumulator obtain- 
ing successive tonal phases. The phase is then 
converted into a waveform using a memory as a 
lookup table. The phase is the input address 
of the memory and the memory word is the wave- 
form value. At the memory output the wave- 
form can be scaled. This is done by multi- 
plying the output by a scaling value. By con- 
trolling this value, decay and envelopes can be 
implemented and also tremlo effects. The cir- 
cuitry can produce many channels by making the 
accumulator a multiword accumulator memory 
and adder. Of course control becomes more 
complicated since the circuitry is shared 
among different channels. In this case all the 

waveforms are summed at the output and this sum 
is converted to an analog voltage in a digital- 
to-analog converter. 

This approach is entirely digital with 
much of the tonal generation process easily 
controlled. The frequency is specified by the 
word in the frequency control RAM; the wave- 
form produced is the waveform in the RAM and 
the amplitude envelope specified by another 
RAM. Thus all of the above can be controlled 
from a microprocessor merely by the processor 
writing into the RAM's. 

This approach is computationally simple. 
An addition is required to generate the phase, 
a memory access to find the waveform value and 
a multiplication to implement envelope func- 
tions. In addition a summation must be per- 
formed over all tones. The approach is flex- 
ible in that all important parameters can 
easily be processor controlled. A logical 
microprocessor function is envelope gene- 
ration since envelopes tend to change slowly. 
The full resources of the microprocessor 
would then be available for this function. 


















— — 









Figure 1. Direct digital svnthesis 



BOX 1579. PALO ALTO CA 94302 

Synthesizer Design Goals 

The design goal was to allow sophisti- 
cated performance yet be of simple design. 
This was attempted by making the hardware 
simple and putting much of the complexity in 
the software. This affords maximum flexi- 
bility since software can be updated and 
changed easily. It also results in simpler 
hardware. The goals of sophisticated per- 
formance and simple design are somewhat exclu- 
sive and some compromises had to be made. 
The compromise was in the number of channels. 
In the approach taken the complexity is some- 
what proportional to the number of channels. 
Thus to simplify the design a small number 
of channels was implemented since this will 
allow lower speed circuitry and some data 
multiplexing can be used. Furthermore 
in taking this approach a modular design can 
be implemented. To gain increased perfor- 
mance more of the identical synthesizer 
modules can be used to increase the number of 
channels. Once the design is completed it 
is easy to duplicate the circuitry. 

An important use of the synthesizer was 
in constructing an electronic organ which 
sounds like a pipe organ. Basically elec- 
tronic organs, especially affordable ones, 
don't sound like pipe organs. Aside from 
the accoustical environment where the organ 
is usually located there are several reasons 
for this. First of all there is usually no 
chorus effect. This is the simultaneous 
sounding of several tones of the same or 
octavely related pitches, each tone 7 sounding 
at a slightly different frequency. 1 J Most 
organs use a small number of oscillators, 
often 12, corresponding to the 12 notes in 
an octave, while pipe organs have the equiv- 
alent of thousands. Furthermore; separate 
manuals often use the same oscillators fur- 
ther aggravating the problem. Another sig- 
nificant problem is that electronic organs 
are limited in timbre or waveform generation. 
Usually the waveforms are constructed by 
passing a signal, such as a sawtooth, through 
a filter representing the desired sound. Also 
the filter usually covers 5 octaves. Coupling 
these two restrictions limits the waveforms 
that can be generated. Third, the transient 
effects of pipe organs are rarely incorporated. 
The pipe organ has definite decay and attack 
characteristics. If these effects are included 
they are usually only rudimentarily implemented 

Since the proposed synthesizer was to 
eliminate as many of the above deficiencies as 
possible the design must implement the effect 
of many oscillators giving a "chorus" effect; 
it must have very flexible waveform generation; 
and it must provide for attack and decay en- 

Initially several constraints were im- 
posed upon the design. Fundamental to the de- 

sign were performance constraints. First, the 
synthesizer must be capable of generating at 
least 25 different tonal frequencies. This 
figure was considered a lower bound since it is 
desirable to produce many more. Since there 
are ten fingers and two feet the maximum num- 
ber of notes that one can play is 12. Having 
a minimum of 25 possible tonal frequencies 
insure that at least two tonal frequencies 
can be generated for each note. Secondly 
there should be a minimum of 4 waveforms or 
tonal sounds generated at one time. Third, 
there must be control of the envelope or ampli- 
tude variations such as trenlo. This control 
must also be programmable. Fourth, the fre- 
quency must be specified to a high degree of 
accuracy. This will allow close frequencies 
to be used for the same note giving a chorus 
effect. This frequency control must also be 
programmable. Finally the word length must be 
at least 12 bits giving a SNR (Signal -to-Noise 
Ratio) of 72 dB. In addition the number of 
samples specifying the waveform should be at 
least 1000. This allows waveforms to be 
accurately specified. 

In addition there were important hardware 
goals. First, no non-standard technology 
should be used. Emitter coupled logic was not 
to be used, nor were special purpose chips 
which are difficult to find. Basically the 
design should use standard off the shelf IC's 
such as TTL and standard MOS. Second, the 
design, especially the controller, should be 
simple. Since the design was to be done in our 
spare time difficult trouble shooting problems 
were to be avoided. In addition a simple 
design insures simple maintenance. Next, the 
design should use a reasonable amount of 
power - about 10 watts maximum. Finally, a 
medium number of IC's, less than 100, should 
be used to fabricate the synthesizer. 

A critical part of the design is the 
software. The software was to be designed 
with certain guidelines. First, to try to 
keep the synthesizer hardware to a minimum, 
software was to be used to its fullest extent. 
This would not only allow for simpler hard- 
ware design but potential performance increase. 
In the future microprocessors will have better 
instruction sets and will be faster. Since a 
large part of the processing will be done by 
the processor there is potential increased 
performance. In addition the microprocessor 
should completely control the synthesizer. 
This would allow for maximum flexibility in 
modifying and changing the synthesized 
sounds produced by the system. In essence 
all the high speed data processing will be 
done by the synthesizer card and all the data 
storage, general bookkeeping and control will 
be done by the microprocessor. 

Finally, since the purpose of the syn- 
thesizer is to generate musical sounds, some 
input device to indicate which sound or notes 



BOX 1579. PALO ALTO CA 94302 

to produce is necessary. The obvious selec- 
tion is a piano or organ keyboard. This 
appears to be the ideal input as the synthe- 
sizer is considered an instrument and should 
be played like one. In addition the inter- 
face must be under interrupt control. Sof * - 
ware polling is not practical since little 
software time would be devoted to other 
synthesizer software functions. 

The synthesizer design method chosen 
was the direct digital method described 
earlier. This offers much in potential per- 
formance and also is computational efficient. 
It further lends itself for integration with 
a standard 8 bit processor. 

Synthesizer Hardware Design 

The tone generate part of the synthe- 
sizer consists of five fundamental parts: 
frequency generator, waveform, weighting, 
accumulation and output, and control (see 
figure 2). The frequency generator determines 
the phase at each sample point. This phase, 
of course depends directly upon the fre- 
quency control word (FCW) provided by the pro- 
cessor. The frequency generator recursively 
generates the newest phase by adding the FCW 
to the prevoius phase. The phase is then 
passed to the waveform section. Here the 
phase is "mapped" into the waveform: from the 
phase, nj |< the waveform, F k (0 n k ), is de- 
termined, (n is the sampling number and k is 
the waveform number.) Then the waveform is 
passed to the weighting section, where a pro- 
cessor controlled value, W k , is used to scale 
the waveform obtaining WkFjjPn^) . Finally 
these values are summed in the'summer section 
obtaining the summed output of each of the 32 
tone generators. The output of course is a 
digital number which is converted into an 
analog voltage. 

As has been described earlier an important 
goal of the design is complete processor con- 
trol over tone generation. In the design the 
processor does have complete control over the 
synthesizer. The processor can write into all 
controlling RAM locations. First the pro- 
cessor controls the frequency of each tonal 
frequency oscillator by writing the frequency 
control word (FCW) into the frequency gene- 
rator RAM. Second, the processor is able to 
select the waveform, f k> by programming the 
waveform number in the waveform select ROM. 
Finally, the processor can select the scaling 
values, W k , for tone generation. Since all of 
the above are processor controlled they can 
change with time. However if a standard 8 
bit processor, such as the 8080, is used up- 
dating all tone generators may be limited to 
approximately once every 2 milliseconds. An- 
other important processor input is the wave- 
form; (F k (0 n>k ). By controlling the above 
parameters and waveforms the processor can 

completely specify the tonal waveform. 

The interface between the synthesizer and 
processor is a memory map interface such as is 
used in the 6800 and the PDP-11. Approxi- 
mately 128 bytes are required to specify the 
FCW, waveform number, and weight. Obviously 
if separate I/O addresses are devoted for each, 
control I/O space will be several reduced. In 
addition the waveform memories are not treated 
as memory space. Each waveform will require 
a IK by 12 bit memory. This could require 2K 
bytes addressing space. With multiple wave- 
forms the waveform memory space can easily 
become very large. Instead all waveforms are 
entered through a common port. As the data 
is entered an internal counter is incremented 
providing the address for the waveform RAM. 
Another port (memory address) is used to zero 
the counter and specify which waveform memory 
is to be loaded. Thus all waveform data is 
transferred to the same location. The inter- 
face circuitry allows the processor to 
write into the various RAM's (FCW, wave- 
form number, weight, and waveform) control- 
ling the synthesizer. 

The controller is a simple circuit con- 
sisting of a ROM and a register. As the ROM 
is sequenced the control waveforms are gene- 
rated. All the timing for the synthesizer 
is derived from this circuitry. 




FCW WRITE -i "_ T~ 











Figure 2. Synthesizer block diagram. 



BOX 1579. PALO ALTO CA 94302 

Frequency Generator 

The frequency generator (see figure 3) 
generates the phase for each of the tona] 
generators. The generator is composed of a 
frequency control word (FCW) RAM, a phase RAM 
and an accumulator. All calculations are per- 
formed using two 8 bit words. Before tone 
generation the processor loads the FCW RAM 
with the two bytes specifying the FCW for 
each tone generator. This value is added to 
the previous tone generator's phase in two 
8 bit additions obtaining the new sample 
period phase. The RAM's used have a cycle 
time of about 250 nsec. Since there must be 
two memory reads and two memory writes it 
takes about 1 microsecond to generate the 
phase. Implementing 32 tonal generators re- 
quires about 32 microseconds resulting in an 
output sampling frequency of about 30 kHz. 
The most significant 10 bits of the phase is 
sent to the waveform section. 




Figure 3. Frequency generator. 


The waveform section (see figure 4) con- 
sists of either a 4K or a 16K word by 12 bit 
memory. There are three different types of 
operations using this memory. The primary 
cycle is the waveform generation cycle. Here 
the phase from the frequency generator is used 
as the memory address and the waveform value 
is read out. The next type of memory cycle 
is a processor write cycle. In this case the 
address comes from the internal word counter 
and the written data conies from the processor 
data bus. In this cycle the processor loads 
the waveform memory. The third type of cycle 
is the refresh cycle. Here a memory read is 
performed with the address coming from the re- 
fresh counter. At the conclusion of this 

cycle the referesh counter is updated. This 
refresh operation occurs every 32 microseconds 
for the 16K memory. 

The memory uses 16 pin 4K or 16K dynamic 
memory chips. These chips require the address 
to be multiplexed in conjunction with two sep- 
arate clocks: row address and column address 
strobe. These clocks are used in multiplex- 
ing the address to the memory chips. Since a 
memory fetch is required by the synthesizer 
once every microsecond and the memory has a 
0.5 microsecond cycle time a spare cycle is 
available every microsecond. The spare cycle 
can be used for either refresh or a processor 
write cycle. The refresh is given precedence. 
If the processor attempts to write data and 
the synthesizer is not ready a memory wait 
will be asserted until the synthesizer is 
available and the cycle is completed. It is 
anticipated the memory maybe loaded under DMA 
control. This will allow the waveform to be 
loaded as the note is initiated. 







V v ir 









Figure h. Waveform section. 


The memory output is sent to the weighting 
section (see figure 2). First the data is con- 
verted from parallel to a 12 bit serial word. 
Then multiplied in the multiplier (25LS14) and 
finally summed in a serial adder (25LS15). 
The serial multiplier and adder have moderate 
speed and low power. Use of the multiplier 
results in a 1.0 microsecond multiply time 
and minimal chip count. 



BOX 1579, PALO ALTO CA 94302 

Processor Interface 

Switch Interface 

As far as the processor is concerned the 
synthesizer is a write only memory. This 
simplifies the design since no read bus 
drivers or multiplexers are required. The 
cycle is started by the processor setting a 
write request flip-flop asserting the memory 
wait line. When the synthesizer has an avail- 
able time slot the write operation is per- 
formed. At the conclusion the request flip- 
flop is cleared and the cycle is completed. 
There are two types of write cycles. The 
first is into the parameter RAM's (210Ts). 
These RAM's specify either frequency, waveform 
number or weighting and have an approximate 
250 nsec cycle time. The other RAM cycle is 
for the dynamic waveform memory. This requires 
a cycle time of approximately 0.5 microseconds. 

Con troller 

The controller consists of a read only 
memory (ROM) and a buffer register. This is 
a very simple controller with the successive 
RUM address determined from the ROM itself. 
As the ROM is cycled it generates the various 
controller signals. Since there is no inputs 
to the controller, other than the 20 MHz 
clock, the operation is straightforward. A 
simplified timing diagram is shown in figure 5. 
As can be seen each cycle consists of two 
major timing parts. In generating the phase, 
first the lower byte is calculated in the 
first half of the cycle. During the second 
half the upper byte of the phase is determined. 
The waveform memory uses this phase during the 
first half cycle to determine the waveform 
value. The last half of the cycle is" avai Table 
for memory refresh or for writing new waveforms. 
The weighting section loads the multiplier when 
the waveform data becomes available - in the 
middle of the cycle. 

















The switch interface (see figure 6) 
allows the keyboard to communicate with the 
processor. Rather than putting switch de- 
bounce circuitry at each switch a common digi- 
tal debounce circuit was constructed. This 
results in much of the complexity of the in- 
terface. The counter continuously counts 
generating addresses corresponding to differ- 
ent switches. At each address a RAM is used 
to remember the status of a switch - essen- 
tially how long it has been on or off. To 
understand the operation first assume the 
switch is off. At this time the RAM will have 
zero stored for that switch. As the switch is 
turned on the status count will increase each 
time the counter reads the switch. Finally a 
switch "on" threshold is reached and that 
switch is considered on. The operation is 
similar when the switch becomes off. Hyste- 
resis has been added, by having the threshold 
when the switch is off greater than when the 
switch is on, to eliminate contact bounce. If 




T-p-i cnj_r , o ^ a S Ar n"tlissi. zs£* "timing'" a 


Figure 6. Switch interface. 



BOX 1579, PALO ALTO CA 94302 

a switch change is detected, that is, the 
switch becomes on or off, an interrupt request 
flip-flop is set. This also stops the counter. 
The processor responds to the interrupt by 
first reading the counter to determine which 
key (switch) is changing. Then the processor 
uses the address to read the RAM and determine 
the switch status. This information is used 
by the software to control the synthesizer 
and will be discussed further in the software 

This may seem excessively complicated but 
it is much simpler than having a hundred or 
so debounce circuits. In addition the same 
switch interface can be used to interface the 
waveform selection switches - stop selectors 
on an organ. 


Introduction. The software to control 
the synthesizer and achieve the synthesis of 
a note can be broken down into four parts. The 
first parts deals with initializing the system 
such as setting up tables, pointers and 
counters. The remaining three parts deal with 
the start of a note, the synthesis of the note 
and termination of the note. The software 
does a great deal of data processing. The 
following paragraphs discusses the software in 
more detail in the hopes that the reader will 
get a better feel for the extent of the soft- 

Initialization . The software has been 
written to accommodate 128 different notes 
(forty more than a piano keyboard and six more 
than two organ keyboards). Information on 
each note must be stored, so that when the 
note is played the software can fetch the in- 
formation from a note table. Table 1 shows 
the information that is stored for each note. 
The first entry, the timbre number, tells the 
software which output waveform the software 
should use to generate the given note. At 
present there are four output waveforms stored 
in the synthesizer card. The attack rate is 
used to determine how long the attack period 
will last. An eight bit register, the attack 
time, is summed with the attack rate every 
clock period. The clock period occurs every 
2 milliseconds. When the software detects a 
carry from the above addition, the software 
will change the state of the note from the 
attack state to the steady state. For 
example, if the attack rate is set at 4, 
the attack period will last 64 clock periods 
or 128 milliseconds. Each note has its own 
rate so the attack period can be different 
for every note. 

The third entry in the note array is the 
attack envelope number. At present there are 
four attack envelope tables stored in software. 

The attack envelope number indicates which 
envelope the software should use for the given 


Note Table 

Each Entry (128 Notes) Includes the 
Following Elements: 

Timbre number 

Attack rate 

Attack envelope number 

Decay rate 

Decay envelope number 

Steady state rate 

Steady state envelope number 

Frequency modulation rate 

Frequency modulation envelope number 

The note table also contains the infor- 
mation for the decay and steady state period. 
This information is used in the same manner 
as the attack information. However, there is 
one difference with the steady state rate. 
The software never tests the carry. This 
means that once the note is in the steady 
state it will remain in that state, except 
for one condition; when the software has de- 
tected that a key is no longer being played, 
the software will change the state of that 
note to the decay state and thus the note will 
be terminated at the end of the decay period. 

The final two entries in the note table 
are the frequency modulation information. 
This information would be used to modulate 
the output waveform. However, at present the 
software has not been written to implement 
frequency modulation. 

Besides the note table there is a status 
active table. The synthesizer can generate 
up to 32 notes simultaneously. The status 
active table contains information on only 
those notes that are being generated. The 

number of entries in the status active table 
depends on the number of notes that are simul- 
taneously being played. If no notes are being 
played, then of course there would be no entries 
in the status active table. The information 
stored in the table for each active note is 
shown on table 2. There are fourteen parameters 
listed for each note. The note number indicates 
which note from the note table is being gene- 
rated. The timbre indicates which output wave- 
form should be used to generate the note. The 
attack rate, attack envelope, decay rate, de- 
cay envelope, stready state rate, steady state 
envelope, frequency modulation rate and fre- 
quency modulation envelope is the same infor- 
mation that is stored in the note table. This 
information is transferred to the status active 
table when a note becomes active. The Fre- 



BOX 1579. PALO ALTO CA 94302 

quency Control Word (FCW) is directly related 
to the note number. The FCW is the actual in- 
formation feed to the synthesizer to indicate 
tonal frequency produced. Following the FCW 
in the table is the FCW synthesizer address. 
This address is the location the FCW is stored 
in the synthesizer. The next entry, amplitude 
synthesizer address, is the location the en- 
velope amplitude data is stored in the synthe- 
sizer. Finally, the timbre number address, 
this address is the location the timbre number 
is stored in the synthesizer. The initili- 
zation software sets up the pointers to the 
status active table and clears the status 
active counter to indicate that all 32 note 
generators are available. 

Status Active Table 

This table contains status information 
on all active notes. Up to 32 entries 
are possible with each entry composed 
of the following: 

Note number 


Attack rate 

Attack envelope 

Decay rate 

Decay envelope 

Steady state rate 

Steady state envelope 

Frequency modulation rate 

Frequency modulation envelope 

Frequency_ Con tro 1 Word ( FCW) 

FCW synthesizer address 

Amplitude synthesizer address 

Timbre number synthesizer address 

The initialization software also has the 
duty of setting up the timbre waveform in the 
synthesizer. The synthesizer card can hold 
four timbre waveforms. Each waveform consists 
of IK 12 bit words which breaks down to 1.5K 
bytes. The waveforms are stored on cassette. 
The software transfers the data from the cas- 
sette to the synthesizer card. There is no 
restriction on the waveforms. They may be a 
simple sine wave or a complex waveform. This 
freedom provides the means of emulating ins- 
trument sounds, since the instrument's wave- 
form can be stored in the synthesizer. 

As mentioned, a note can be in one of 
three states: attack, steady state or decay. 
A state table is used to indicate which state 
a note is in. There is a maximum of 32 en- 
tries in the state table. Table 3 shows the 
information contained for each entry. 

State Table 

This is the operating array for 
each active note. Each entry has 
the following elements. 




Envelope base address 

Amplitude synthesizer address 

Note number 

The first entry is the state. This en- 
try indicates which state the note is in. 
The second entry is the time. This entry in- 
dicates how long the note has been in the 
given state. Time is also a pointer to the 
location in the envelope table the note is 
presently using. The location contains the 
amplitude datum for the output waveform. 
Every clock period the rate is summed with 
time to give a new time and a new address for 
amplitude datum. If a carry is detected the 
software will change the state of the note to 
the next state. However, there is one problem, 
you do aot want to leave the steady state and 
start the decay until the note has been re- 
leased. To prevent this from occurring the 
software does not test the carry from the 
steady state time and rate summation. Thus 
the note will not go into the decay state 

until the software detects that the note has 
been released and chages the state of the note. 

The next entry in the table is the envelope 

base address.. This, is the address -of the first 

entry of the envelope table for the given state. 
By adding time with the base address, the lo- 
cation of amplitude datum is determined. This 
datum is loaded into the synthesizer amplitude 

The final storage table is the note fre- 
quency table. This table contains the frequency 
control word for each note. Each FCW consists 
of two bytes and since there are 128 notes, the 
table occupies 256 bytes. The initilization 
software loads the proper FCW for each note. 

Besides the above tables, the initiali- 
zation software sets up two stacks. One stack 
contains the addresses of the available gene- 
rators. The software loads the 32 addresses of 
the status active table. The second stack con- 
tains the addresses of the generators that are 
active. Since there are no active generators 
to start off, the software clears this stack and 
and clears the counter which indicates how many 
notes are active. 

The final requirement of the initialization 
software is to set up pointers to the starting 
addresses of the above tables and to set up 
the numerous counters that are used. 



BOX 1579, PALO ALTO CA 94302 

Note Detection . As mentioned earlier the 
software excluding initilization can be broken 
down into three sections: note detection, 
clock detection and note deletion. The note 
detection software is used when a note is 
first detected. When a note is first played, 
the keyboard hardware detects the note and 
causes an interrupt. The interrupt vectors the 
processor to the note detection software. The 
first thing the software does is obtain the note 
number from the keyboard hardware. Using the 
note number the software determines the Fre- 
quency Control Word and loads it into the status 
active array. Next the note parameters are 
moved from the note table to the status active 
array. The synthesizer addresses are moved to 
the status active table. The state table is 
set up with the attack parameters. The Fre- 
quency Control Word and timbre number are 
loaded into the synthesizer. The output ampli- 
tude is set to zero. The generator available 
pointer is incremented to point at the next 
available generator. If no generators are 
available, the generator full flag is set. 

The note detect software does not start 
the synthesis of a note but merely sets up the 
data so that it may begin. It is during the 
clock interrupt software that a output from 
the synthesizer is generated. 

Clock Detect . Every 2 milliseconds the 
software is interrupted and vectored to the 
clock detect software. The software will then 
cycle through the active notes. For each note 
the time datum and rate are summed to give the 
address of the amplitude data. If a carry is 
detected the software will change the state of ' 
the note. The amplitude data is transferred 
to the synthesizer card. If the software de- 
tects the end of the decay state then the soft- 
ware will delete the note from the status active 
table and the state table. It will decrement 
the generate available count to indicate that a 
generator has been freed. 

amplitude of each note can be individual con- 
trolled. The flexibility of the design is do 
to the fact that much of the processing and con- 
trol is in software. 


R. B. Cotton, "Tempered Scale Generation 
From a Single Frequency Source," Journal 
of the Audio Etiaineerina Society* VoT. 20, 
pp. 376-382, June 1972." 

2. J. A. Moorer, "Signal Processing Aspects of 
Computer Music: A Survey," Proceeding of 
the IEEE, vol. 65, no. 4, pp. 1108-1137, 
Aug 1977. 

Note Delete . When the keyboard hardware 
detects the end of a note, an interrupt is 
sent to the processor. This interrupt cause 
the processor to vector to the note delete 
software. The software will obtain the note 
number from the hardware. It will go to the 
state array and change the state of the note 
to decay state. The clock detect software de- 
tects the end of the decay state and does the 
necessary bookkeeping as mentioned above. 


The synthesizer described in this paper is 
intended to provide high quality musical syn- 
thesis. Up to 32 simultaneous notes can be 
synthesized. The attack, decay and steady state 

3. J. M Chowning, "The Synthesis of Complex 
Audio Spectra by Means of Frequency Modu- 
lation," Journal of the Audio Engineering 
Society, Vol. 21, no. 7, September 1973. 

4. J. Tierney, C. M. Rader and B. Gold, "A 
Digital Frequency Synthesizer," IEEE Trans. 
Audio Electroaccoust., vol. AU-19, pp. 48- 
56, March 1971. 

5. A. Heaberlin, U.S. Patent No. 4,003,003, 
"Multichannel Digital Synthesizer and 
Modulator," January 11, 1977. 

6. J. Snell, "Design of a Digital Oscillator 
Which Will Generate up to 256 Low Distor- 
tion Sine Waves in Real Time," Computer 
Music Journal, vol. 1, no. 2, pp. 4-25,1977 

7. Richard H. Dorf, "Electronic Musical Instru- 
ments," Radiofile, New York, 1968. 

8. A. Popoulis, "The Fourier Inteqral and Its 
Applications," McGraw-Hill, New York, 1962. 


To give those who are mathematically in- 
clined a better feel of the mechanics of the 
synthesizer, the following is a brief mathe- 
matical treatment of the synthesizer tone 
generation. The section consists of two 
parts. The first describes a mathematical 
representation of the tone generation pro- 
cess. The other section describes the errors 
introduced into the tone generation and shows 
how they can be determined. 

Assume we are interested in generating a 
periodic waveform, f p (t), which has period, 
Tp. In addition let us generate a function 
fs(«0 corresponding to a single cycle of the 
waveform but with period, 1: 



BOX 1579, PALO ALTO CA 94302 

f s (t/ y = f P (t) for °- t<T P - 

In generating this waveform we will generate a 
sampled representation of the signal with sanple 
period, T s . Thus we are only interested in 
fp(t) at discrete points/ specifically: 

t = nTg for — » < n < « . 

Since the waveform is periodic we are only in- 
terested in the "phase," not the integer number 
of cycles. This "phase," normalized between 
and 1, has zero representing the start of the 
cycle and one representing the end of the cycle. 

We can define a remainder function, REM, 
which gives us the normalized phase. Mathe- 
matical we first separate nT s into the integer 
number of cycles, K, and the phase at the nth 
sample period . 

nT s - KT p + »n T P 

This is done by finding K, the integer part of 
(nT s /Tp). Then 

K - K " KT p)/ T p 



"n ■ rem < n vy- 

<«n> * V nT s> 

We want a recursive method of generating 
f D (nT § ) or equivalent f s t0 n ). We can generate 
fs(0n/ by generating 0n recursively. We can 
generate n+ -| from n : 

J n+1 = REM (nT s /T p + T s /T p ) 
= REM (k + n + T s /T p ) 
= REM (0 n + T s /T p ) 


Thus we can recursively generate f s (0n) or 
equivalently f p (nT s ) by recursively generating 
n from (1) and then determine f s (gS n ) from p . 

Equation (1) is easily implemented by 
using a digital register representing n , and 
an adder. The remainder function is easily 
implemented by truncating the register output. 
This register and adder are often called the 
phase accumulation register. At each sample 
period T s /T p is added to the previous phase de- 
taining the new phase. This phase, n , is then 
converted into f s (0 n ) by table lookup. A RAM 
or ROM is used to represent t s (0 n ). 

Error Analysis 

If Mrfn) = fp(nT s ) with n = REM(nT<;/T p ) 
for all nt s then there will be no distortion 
provided, of course, the Nyquist criteria is 
meet: the highest frequency of f p (t) must be 
less than one half the sampling frequency, 
1/T S . 

However fp(nT § ) will not equal fs(# n ! for 
two reasons. First the waveform fs(*0 w ""^ 
have finite entries and will appear as in 
figure 7. This is because f s (0n) is constant 
between entries: in practice trie phase is 
truncated and the next lowest phase, say 0-, is 
is used to approximate f s (0 n ). Thus f s (0n) is 
generated - not fp(nT s ) and we can analyze the 
spectral content of this waveform to determine 
the distortion. The above introduces error 
when the phase has more resolution than the 
table used to represent f s (0n)- Of course 
if the phase, n , is not truncated than there 
are no errors introduced. But typically the 
phase is generated with more resolution and 
will be truncated. The other source of error 
is quantization error. Since the word length 
used to specify the sample value is finite, 
errors are introduced. This results in white 
noise (constant noise power density as a 
function of frequency) and can be controlled 
by making the word length long enough. 12 
bits gives a signal-to-noise ratio of 72 dB 
and is probably adequate. First the spectrum 
of f s (0 n ) approximating the waveform f p (t ) 
will be determined. A single cycle of the 
periodic waveform is sampled for storage in 
the table or memory (see figure 7). The 
sampling interval is the period divided by N, 
the number of entries in the waveform table. 
TM waveform is. sampled at each point and then 
this value is held constant until the next 
point (flat-top reconstruction). Thus the wave- 
form in the memory, 

has the following re- 

Ut) = 



h(t-n/N) f s (n/N) 

h(t) = 


o< t<l/N 

The tone generation process essentially deter- 
mines a phase and then uses this to lookup the 
waveform value in the memory. This waveform 
determining process can be separated into 
two equivalent processes. First the waveform 
will be produced from the ROM at the correct 
frequency with infinite precision. Next this 
signal will be sampled at the sampling rate, 
1/T S . This is identical to using the phase 
directly to determine the waveform value. 



BOX 1579, PALO ALTO CA 94302 

f U) 


-f (t/T ) 
P s 


k. %v<^"»r -r- 

Figure 8. 

harmonic error 

Spectrum of memory waveform 

Figure 7- Memory 

waveform, f (^) 

First the waveform from the ROM at the 
correct frequency will be a periodic signal 
identifical to a single cycle of the memory 
waveform with frequency 1/(T S /T p ). This wave- 
form has the following representation: 

f(t) = Z h((t-n/N)T p ) 

f w (nT s ) 

Here f is a periodic signal with f (nT ) = 
f s (REMTnT $ /Tp). This has the following spec- 
trum (see figure 8): 

F(u>) = 


F w (o) -na) s ) 

Obviously the waveform is not bandlimitea. 
There are abrupt transistions between entry 
points. This waveform will then be sampled at 
the sampling rate, T Let us further assume 
the signal we are generating is a sinewave as 
has been represented in figures 7-8. This isn't 
necessary for error analysis but it does allow 
us to separate harmonic error from other errors. 
Essentially all spurious frequencies above h 
the sampling rate are translated into the region 
-T s /2 to T~/2 (see figure 9). Thus we can 
separate the error into two parts: harmonic and 
non harmonic. All harmonic error less than 
Kyquist bandwidth will remain undistorted. 
Since this error is harmonic it will not be as 
important as the non-harmonic error and will be 
disregarded here. However frequencies above 
the Nqyuist bandwidth will appear as spurs. 
The waveform to be generated can be separated 
into harmonics or partial s. The fundamental 
will have one cycle stored in the memory thus 
N will be 1024 because all memory locations are 
used to specify the cycle. However for the 
other partial s more cycles will be present in 
the memory and the effective N will different. 
For instance, for the first harmonic two cycles 
will be represented in the memory. Therefore 
the effective N is 512. A procedure for de- 

-Fundamental (l/T ) 






(3/(NT )-l/T ) will 

P P 

produce spur at 

(3/(NT )-l/T )-l/T 
p p s 





Spur produced 

Figure 9. Spectrum of generated signal. 

termini ng the non-harmonic errors is as 
follows: For each partial, determine the 
effective N. This will be 1024 divided by 
the harmonic number plus 1. Next determine the 
first frequency above the Nyquist bandwidth as 
above. This will be the strongest spur and 
all frequencies above this will be spurious. 
Repeat this for each partial in the generated 
signal. In addition the spurs are proportional 
to the strength of the corresponding partial. 
Two possible approaches to determine allowable 
error could be considered here. One might be 
to consider just the maximum spur. But pro- 
bably a better error function is to determine 
the total spurious error. This will be a power 
summation over all spurs and over all partial s. 
Using this for a particular waveform and fre- 
quency the signal-to-noise ratio can be de- 



BOX 1579, PALO ALTO CA 94302 


Byron D. Wagner 
1701 Viewmont Dr. 
Los Angeles, Ca. 90069 


This paper describes the relative 
ease with which an inexpensive person- 
al computer can be configured into a 
highly sophisticated, personalized 
tool or test instrument with perfor- 
mance equaling or exceeding that of 
costly, commercially available, less 
flexible equipment. The author 
details the design, construction, 
software development, smoke-test- 
ing and calibration of such a device 
- a real-time, 1/3 octave audio 
spectrum analyzer using a color 
television as a graphic display 
screen. Applications for this part- 
icular device are discussed and 
guidelines are offered for custom- 
izing systems to the users needs, 
either as a standalone instrument 
or as an automated controller to be 
integrated into an existing unit. 
Examples suggested include: mixing 
consoles, theatrical lighting boar- 
ds, and a scanning electron micro- 


According to some scientists, 

the- .^aixi-lity to create artificial 

extensions of fingers, specialized 
for different purposes, is what 
originally separated cavemen from 
the animals. History is split into 
periods representative of the mat- 
erials used to form the tools, ie. 
the Stone Age, the Bronze Age. As 
man grew more sophisticated, so 
did his tools. Bits of sharpen- 
ed rock and bone for cutting and 
grinding became knives and spear 
points, later arrows. With the 
invention of the wheel, man was 
no longer totally dependent on 
beasts of burden, and transpor- 
tation became more efficient. 

By the Middle Ages, technol- 
ogy and craftsmanship had raised 
the original utilitarian cutting 
tool to the level of being a power- 
ful symbol. Ornately jeweled swords 
and scepters conveyed authority, 
items of ritual never put to prac- 
tical use. For the most part, how- 

ever, the advance of technology was 
not an end in itself, but a means of 
gaining leverage over materials used 
in making the physical world a better 
and safer place to live. The fact that 
people could derive pleasure from the 
esthetic challenges of organization is 
displayed in the creative patterns 
commonly found in early pottery and 
weaving, with even the most common 
household items crossing the line 
between strict functionality and the 
communication that is art. 

When I was young, my father took 
me to visit a craftsman who made his 
living by repairing musical instru- 
ments, expecially brass horns. As I 
was admiring the amazing assortment of 
metal-working implements, I asked where 
someone might buy such unusual, special- 
ized equipment. "I wouldn't know", he 
replied, "Imake my own myself", and he 
smiled with pride. 

People have always been envious of 
such highly personalized and custom- 
ized tools. From the gunfighter's 
pearl-handled revolver (with the hammer 
spur filed off to facilitate quick 
draws), to the pool hustler's weighted 
and balanced pool cue, or the profess- 
ional musician's instrument, voiced 
just to his liking. In each case, the 
modifications are justifiably practical 
necessities for the pro, but luxuries 
to the layman (novice). 

Personal Computers - Getting Beyond 
"Other People's Programs" 

The personal computer provides the 
opportunity to create devices whose 
complexity and degree of personalization 
is limited solely by the imagination 
and motivation of the user. It is an 
accessible alternative for the solution 
of problems. In other words, one can 
trade an abundance of time for a non- 
abundance of money. With the additional 
advantage that, once a written logic 
framework or program is established, 
it can be easily adapted for later uses. 
Organizing information utilizes the same 
techniques whether the data manipulated 



BOX 1579, PALO ALTO CA 94302 















esent s 
h reco 
s , or 
t ing c 
t on t 
e syst 

and a 
as di 
as us 
tant ia 

a stamp 
rds , re c 
an amate 

00 rdinat 
he mass 
em data 

t echniqu 
pplied t 
ment of 
eful as 

1 saving 

ur a 
es . 
es c 
o ho 
ref e 
s or 
s tor 
s in 

, ham 
s t rono 

ry pri 

s t ora 
an be 
me sys 

ing th 
and of 


n, phono- 
radio con- 
at this 
ce curve , 
ge and 
es is al- 
e refer- 
f ers a 

Even though the mass marketplace 
for personal computing products is 
still relatively small in relation 
to other hobbies, the array of 
existing building blocks is pheno- 
menal. Eight-and sixteen-bit CPU's 
cheap RAM and ROM memory, inexpensive 
tape and disc storage, terminals, 
quality hard-copy printers, medium 
resolution black-and-white or color 
displays, speech generation and 
recognition devices, communication 
modems and touch-tone transceivers, 
remote control of household appli- 
ances using AC lines, music syn- 
thesizers, analog to digital and 
digital to analog converters, digital 
logic analyzers, digital frequency 
counter and digital multimeter sub- 
systems, realtime clocks, and even 
medium resolution video digitizers 
exist, most compatible on a "plugin 
and go" basis. Yet perhaps a greater 
benefit is the willingness on the 
part of most manufacturers to pro- 
vide not only prewritten software, 
but also technical support. This 
includes schematic diagrams and 
"theory of operation" information 
on a nuts and bolts level (as 
opposed to the traditional "once 
you buy it we don't want to hear 
from you" supplier-consumer re- 
lationship . ) 

Real Time 

Part of the precision with which 
we operate our hands and fingers 
stems from the fact that they are 
able to send feedback signals to the 
brain that controls their movements. 
This is a classic example of a 
closed-loop servo system (as oppo- 
sed to blindly applied brute force.) 
The system can be disturbed, though, 
if a delay is introduced either 
between the sensing apparatus and 
the controller, or the controller 
and the actuators. With children 

or adults, the more quickly rein- 
forcement follows action, the 
faster the learning process is 
accomplished. This builds the pattern 
of challenge, motivation, and reward 
into a constructive "vicious circle." 
A significant benefit of this behavior 
is that more good ideas will be follow- 
ed through to completion, instead of 
being abandoned because of petty but 
time-consuming obstacles. Studies in 
human ergonometric design confirm 
practical experience relative to the 
speed and accuracy with which a person 
can assimilate raw data. Computer hex 
codes are harder to understand than 
conversational English, which in turn 
is more difficult to read than small 
strings of numbers. The simplest way 
to communicate rapidly comprehensible 
information reflecting changing signal 
parameters is through the use of analog 
position indicator movements referenc- 
ed to a fixed linear or logarithmic 
scale. Convention in the past dictated 
that separate functions required sep- 
arate displays, whether visually 
oriented (meters and gauges) or audio 
(buzzers, bells, and chimes). But 
experience has shown that human beings 
are capable of deciphering numerous, 
although subtle, details from complex 
integrated or holistic sets of stimuli. 
An experienced mechanic listening to a 
car's engine can diagnose many specific 
different ailments relating to carbu- 
ration, valve, timing, and mechanical 
balance problems in much the same way 
as a physician examines a patient. 

An example 
applied to el 
may be found 
and even some 
When a standa 
stereo progra 
channel going 
cuits and the 
horizontal di 
"scrambled eg 
This pattern 
amplitude , s t 
phase, freque 
characterist i 

of this type of design 
ectronic test instruments 
in many recording studios 

home hi-fi systems, 
rd oscilloscope is fed 
m material, with the left 

to vertical drive cir- 

right channel feeding 
rve circuits, a complex 
g" pattern appears, 
simultaneously shows peak 
ereo separation, relative 
ncy, and dynamic envelope 
cs for both channels. 

It has become feasible to design 
instantly reconf igurable output inter- 
faces with ultimte flexibility without 
giving up the capability of providing 
the ultimate in specialization and 
detail. Such devices can be person- 
alized to the idiosyncracies of an 
individual (whether the user is left- 
handed, speaks a foreigh language, 
is male, female, or even a child or 



BOX 1579. PALO ALTO CA 94302 

animal), with different pre-programmed 
levels of complexity. Obviously, this 
is totally opposite to traditional 
instrument design practices and does 
offer disadvantages. These include 
the lack of standardization of oper- 
ation and maintenance procedures. 
However, standards will probably emerge 
Hopefully, they will be generated on 
the basis of need and logic, instead 
of manufacturing expediency and 
economy . 

Obviously, a 
analysis tool sue 
describing, when 
element for a pro 
is capable of ere 
structure suitabl 
mation or indepen 
manual assistance 
even complete ove 
possible while st 
housekeeping func 
trim logic to all 
much or as little 
desired. A good 
would be an autop 
could coordinate 
the direction hea 
dance functions h 
the pilot. 

powerful s 
h as we ha 
used as th 
cess-cont r 
ating a la 
e for comp 
dend actio 
, supervis 
rride by a 
ill retain 
tions and 
ow the ope 
example of 
ilot on a 
turns, eve 
ding and o 
ad been di 

ve been 
e feedback 
ol device , 

lete auto- 
n. Yet 
ion, or 

human is 
ing the 
automat ic 
rator as 
il ity as 

plane that 
n though 
ther gui- 
sabled by 

A Real-World Example 

In the field of professional audio, 
historically, control of signals has 
been limited primarily to the mani- 
pulation of the amplitude of informat- 
ion derived from re_a.l-w.o.r 1 d. .so:urc£s. 
Typically, these are musical instru- 
ments, voices, or in the case of motion 
pictures, and television, sound effects 
or background sounds. In recent years 
this has been expanded to include equal- 
ization (still a function of amplitude), 
artifical reverberation and echo, multi- 
track recording, phasing and digital 
delay lines (time), and most recently, 
pitch or frequency (made possible by 
the use of analog to digital and digital 
to analog conversions). Attempts at 
either the direct synthesis or modifi- 
cations of complex waveforms were limi- 
ted for the most part to electro-mech- 
anical devices (Hammond organ tone 
wheels and Leslie rotating loudspeakers) 
or prohibitavely time-or money- consum- 
ing procedures using large-scale com- 
puters (Music V). Additionally, the 
computer programs did not operate 
interactively in real-time. This state 
of affairs explains why commonly used 
indicators throughout the industry in 
broadcast stations, recording studios, 

and communication networks were 
generally VU or Peak Program meters, 
to be used in conjunction with loud- 
speakers and sometimes an occasional 

Things, however, are changing fast. 
The changes result from such factors 
as the consumerizat ion and mass- 
marketing of voltage-controlled 
synthesizers like the Moog and ARP , 
the skyrocketing technology of large- 
scale integration, and the continu- 
ing demand for more control and real- 
ism in audio recording and sound 
reinforcement. These changes also 
brought about the need for more 
sophisticated measurement tools at 
affordable prices. The ability for 
a human to tell the difference between 
musical instruments, for example, a 
flute or a clarinet, is due to the 
difference in harmonic structures 
and dynamic envelope changes that 
occur during the duration of a note. 
These are determined by the mechanics 
of vibration peculiar to the gener- 
ation of the sound emitted by that 
particular instrument. In the case 
of the flute, a vibrating column of 
air; in a clarinet, vibrating reeds 

In response to this demand, re- 
searchers developed tools like the 
spectrograph, a device which splits 
audio signals occurring in the 
spectrum between twenty and twenty 
thousand vibrations per second into 
bands, much as a prism devides white 
light into its --c-ompon-ea-t- parts « This 
was accomplished by passing complex 
signals through a turnable filter, 
plotting the results, and repeating 
in an overlay fashion after retun- 
ing the filter. This yields a result 
similar to the graphs popularly 
known as "voice-prints". Early 
attempts at real-time analysis were 
cumbersome due to the number of 
filters and meters necessary for 
reasonably discrete identification 
of separate harmonics. If the spec- 
trum is divided into octave bands, 
the required number of filters is ten. 
If half-octave, nineteen; third- 
octave, twnety-seven. That meant 
dozens of tubes and lots of heat. 
In early 19 7 1, Hewlett-Packard in con- 
junction with Altec-Lansing, intro- 
duced an audio spectrum analyzer in 
a relatively small, rack-mount case 
with a CRT display, selling at the 
breakthrough price of approximately 
$3,000. This was followed a few 
years later by a unit from Amber 



BOX 1579. PALO ALTO CA 94302 

Electro Designs of Montreal. It the most attractive points was that 
features an LED matrix display, with all the hardware, with the exception 
dimensions of ten by ten, and retails of filter banks and input signal con- 
for about $2,000. (See Figure #1) The ditioning circuit ry, was available off 
Ivie Corporation, based in Utah, is the shelf. Final choices included : 
now marketing a palm-sized, hand-held, an IMSAI mainframe with front panel 
audio spectrum analyzer with built-in and CPU, Seals and IMS 8K memory 
LED matrix display and calibrated cards, a Sony color monitor* a Byte- 
microphone. This sells for less than saver card, three D+7A's (analog 
half a thousand dollars. to digital converters) and a Dazzler 

color video display interface, all 

As the availability of such devices from Cromemco. (Fig. #2) 
grew, so did their uses. eg.: the study 

of acoustics and musical instrument tone The display format chosen was 

generation, bandpass characteristics of that of vertical, multi-colored bars, 

amplifiers, tape machines, and so on, whose heights rise and fall with the 

and the practice of tuning or "voic- output of the corresponding filter/ 

ing" music reproduction or sound rein- peak detector combination. The 

forcement systems using graphic equal- basic algorithm for translating 

izers to compensate for room charac- analog input amplitude changes to 

teristics, and displays of power band- corresponding changes in bar-graph 

width in disc mastering chains. height was developed, and a test 

program written in basic, along with 

The Hardware a routine that erased the screen at 

the onset of operation. Simply 
In the course of designing a pro- stated, a scanning pattern is estab- 
fessional recording facility with lished and, for a particular bar, a 
built-in video tie-lines connecting the value is input and compared with the 
performing, control-room and playback present position on the screen. If 
areas, the idea of including such an the screen value is greater than the 
analyzer with the ability to distri- input value, a block of color is 
bute a display to the video monitor written at that position. If not, 
in the control room (saving valuable the background color, in this case 
space) coupled with the possibility black, is written into the position, 
of providing the musicians with (See Fig. #3) With the test program 
immediate feedback corresponding to running under Altair, 3.1 Basic at 
their dynamics, proved irresistable . standard processor speed, it took 
Unfortunately, none of the commer- approximately twenty seconds just 
cially available units offered such to write one screen featuring 8 bars, 
a capability without the need for Since the minimum update speed nec- 
some kind of scan conversion, for essary for smooth and natural step- 
compatibility with NTSC video signal less transitions between frames is 
standards. A little investigation about a thirtieth of a second, it 
and guesst imating led to a potentially was both necessary and desirable to 
feasible design for accomplising implement the program in machine 
not only these goals, but a host of language. An early incarnation of 
others: color display, programmable the assembly listing for such a 
overload threshold, with a provision machine language program (still 
for matching existing disk cutter- displaying only 8 bars) is shown 
head curves, the ability to freeze a in Fig. #4. 
frame of display, or display frozen 

frames sequentially for a slow- Study will reveal the relatively 

motion effect (in forward and rever- crude nature of the counting loops 

se) to facilitate the analysis of and the fact that the entire bar is 

signals for music and voice syn- repainted each scan. The quadrant 

thesis and recognition, and the positioning and jumping arithmetic 

ability to display a large number was made necessary by the architec- 

of fullband signals from multiple- ture of the display used - a Cromemco 

channel tape machines or mixing Dazzler (high resolution color mode), 

consoles, to eliminate the head Any memory mapped video display would 

swiveling (and resulting neckache) work equally well. The software was 

from trying to read thirty-two VU expanded and further refined to allow 

meters at the same time. One of interaction with simple switches so 

that a terminal would not be necess- 


ary in the case of dedicated operation 
Hardware bugs were chased down and 
exterminated and PROMS containing the 
final version of software were burned. 
The device was calibrated using the 
time-honored tradition of applying 
signals of known frequency and ampli- 
tude while making the corresponding 
necessary adjustments for proper 
indicat ion . 

success of the project with sugges- 
tions or construction assistance. 
Notably: Mr. Rick George, Mr. Mike 
Ronstadt, Mr. Gale Lester, as well 
as Ms Sally Grimm and Ms Annie 
Moss whose typing tootsies made this 
paper possible, and a special thanks 
to Mr. Allen Immerman for his fear- 
less and tireless help in toggling 
the beast into submission. 

The resulting display is a 
tremendiously useful device - although 
the beauty of its fascinating patterns 
set to music may lead the more cyncial 
to cast aspersions on the purity of 
pragmatic intent of its creator. By 
plugging in other types of sensors, 
eg.: thermocouples, optical position 
encoders, anemometers, medical elec- 
trodes; and changing the reference 
scale and label overlay data, vir- 
tually any type of measurement para- 
meters can be accomodated. The 
dynamic range, resolution, response 
time and ballistics, averaging 
criteria, linearizarion or log scal- 
ing and weighting, can all be defined 
by software. With the addition of a 
modem, remote polling and data acqu- 
isition become possible; with a mass 
storage device (disc, tape, bubble 
memory) the opportunity for exten- 
sive signal analysis and hard copy 
plots. Yet with all this flexibil- 
ity, the system can satisfy the most 
subtle demand for highly specialized 
applications. If time permits, the 
author will discuss interfaces with 
control elements in theatrical 
lighting systems and the automation 
of recording studio mixing consoles 
and electronic music synthesizers 
as well as a scanning electron micro- 
scope function control and stage 
posit ioner . 

About The Author 

Byron D. Wagner, 26, is originally 
from Omaha, Nebraska. He is currently 
employed in the Los Angeles area as 
an independent record producer and 
recording engineer - in addition to 
functioning as a consulting engineer 
for recording studio design, constru- 
ction, and installation. Past and 
current clients include: Carole King, 
Linda Ronstadt, Peter Asher, Music 
Recorders, Inc., Motown Records, 
Ike & Tina Turner, Steve McQueen and 
Ali McGraw. His professional 
affiliations include: The National 
Academy of Recording Arts and Sciences, 
The Audio Engineering Society, the 
SMPTE, the Magic Castle, and the 
American Federation of Television 
and Radio Artists. His still photo- 
graphy has been featured on several 
album covers and he is the host of 
the PBS television program "Singer/ 
Songwriter". His education was 
assembled through such diverse 
facilities as the Omaha public school 
system, Ohio University, Brigham 
Young University, and the Eastman 
School of Music. (And he wants to be 
a movie star when he grows up.) 

Conclus ions 

In view of the current mind- 
boggling leaps in technology and 
the complexity of available hard- 
ware, it is comforting to consider 
that, far from being closed, the 
available options for designing and 
molding a personalized, custom- fit ted 
set of tools with which to create and 
communicate are as rewarding and open 
as in the past, if not more so. 


The author would like to thank 
all those who contributed to the 



BOX 1579, PALO ALTO CA 94302 




>^<rT> L^ INPUT V^ 






-15v _ 


filter; j 




? MEM A 






10 x X) 







LjT 1".". _"■ Epl 







BOX 1579, PALO ALTO CA 94302 






BOX 1579. PALO ALTO CA 94302 

10 OUT 14, 144 
20 OUT 15,16 

30 B = 

31 Z=8192 

32 POKE Z,B 

33 Z=Z+1 

34 IF Z 8 70 5 GO TO 32 
40 C=51 

50 P=8192 

60 1=25 

7 GO SUB 50 

80 C=85 

90 P=8194 

100 1=26 

110 GO SUB 5 00 

120 C=l 19 

130 P=8196 

140 1=27 

150 GO SUB 500 

160 C=153 

170 P=8198 

180 1=20 

19 GO SUB 500 

200 C=187 

210 P=8200 

220 1=29 

230 GO SUB 500 

240 C=221 

250 P=8202 

260 1=30 

270 GO SUB 500 

280 C=255 

290 P=8204 

300 1=31 

310 GO SUB 500 

320 C=51 

330 P=8206 

340 1=33 

350 GO SUB 500 

360 GO TO 40 

500 D=INP(I) 

510 L=C8704-P) 


530 IF (D"2). =L THEN POKE P,C: TO GO 550 

540 POKE P,B 

550 LET P=P+16 

560 GO TO 510 




BOX 1579. PALO ALTO CA 94302 

ASSM 0700 

0700 31 


07 U 



0703 3E 




0705 D3 




0707 3E 




0709 D3 




070B 01 





070E 21 





0711 36 




0713 OB 



0714 23 



0715 78 



0716 FE 




0718 C2 





071B 79 



071C FE 




071E C2 





0721 06 




0723 DB 




0725 21 





0728 CD 





072B 06 




072D DB 



IN 26 

072F 21 





0732 CD 





0735 06 



BARC MVI B, 119 

0737 DB 



IN 27 

0739 21 




CPLAC LXI H, 1004H (" ") 

073C CD 





073F 06 



BARD MVI B, 163 

0741 DB 



IN 28 

0743 21 




DPLAC LXI H, 1006H (" ") 

0746 CD 





0749 06 



BARE MVI B, 18 7D 

074B DB 



IN 29 

074D 21 




EPLAC LXI H* 1008H (" ") 

0750 CD 





0753 06 



BARF MVI B.,221 

0755" DB 



IN 30 

0757 21 




FPLAC LXI H, 100AH (" ") 

075A CD 





075D 06 



BARG MVI B.,255 

075F DB 



IN 31 

0761 21 




GPLAC LXI H* 100CH (" ") 

0764 CD 





0767 06 




0769 DB 


_ J3A6Q 

. LN .33 __... 

076B 21 




HPLAC LXI H* 10 0EH (" ") 

076E CD 





0771 C3 





0774 OE 




0776 11 





0779 00 



077A 00 



077B 00 



077C 00 



077D B9 



077E 00 



077F 00 



0780 00 



0781 00 


nop ,- rtAL 


0782 DA 




0785 70 



0786 19 

0787 OD 



0788 OD 



0789 OD 



078A OD 



78B F8 



078C C3 





078F 36 




0791 C3 





EXEC 0700 


FREAQ 0100 





BOX 1579. PALO ALTO CA 94302 

copyright 1978 DBH 
David B. Harrison, Esq. 
Owen, Wicker sham & Erickson 
433 California St., San Francisco, CA 94104 
(415) 781-6361 


A cursory view of the patent 
system with focus upon its appli- 
cation to personal computing. The 
difference between inventions 
and patents, and requirements for 
patentability including a general 
discussion and update on the pat- 
entability of software; the rela- 
tionship of patents to copyrights 
and trade secrets; and, obtaining 
licensing and enforcing patents — 
why bother? 


When you stop to think about 
it, we live in a truly exciting 
time. And this fact is proven, 
at least, by our common interests 
in the present and future of 
personal computing which bring 
us together at this Second West 
Coast Computer Faire — note 
please that the word "faire" is 
spelled F-A-I-R-E. In Marin 
County where I live, in late sum- 
mer each year there is a Ren- 
aissance Pleasure Faire, spelled 
F-A-I-R-E, which celebrates the 
Renaissance, that glorious period 
of growth in the arts and sciences 
of four hundred years ago. Today, 
we celebrate a new Renaissance, 
a golden age of computer power and 
promise, an advanced technology 
which we have inherited from all 
of the innovations of the past, 
and yet which is but a crude and 
primitive glimpse and promise of 
what we can expect for the future. 

Over the last three hundred 
years it has come to be recognized 
by enlightened people and their 
governments that human progress 
is promoted, particularly in the 
useful arts, by having laws which 
protect creative works of author- 
ship and invention. What I'm 
talking about are the copyright 
and patent laws which protect 
authors and inventors.! Now, 
by laws which protect authors 
and inventors, I mean laws which 
shield them from infringements 
by others. I do not mean laws 

which grant inventions to inven- 
tors or works of authorship to authors 
- — they own their original works from 
the moment of creation. 

What our intellectual pro- 
perty laws do is to recognize 
that unless a reasonable legal 
shield is provided, authors and 
inventors simply will not be 
motivated to disclose their works 
and discoveries — they will 
keep them secret, and the rest of 
us will not learn about them and 
have a basis upon which to make 
improvements. Leonardo Da Vinci's 
secret discoveries or speculations 
in the Fifteenth Century about the 
parachute and the helicopter2 were 
uncovered and appreciated only in 
recent years, and provide a good 
example of the point I am trying 
to make. Fifteenth Century laws 
did not promote the useful arts 
and offered no incentive to men 
and women like Leonardo to make 
their discoveries public. 

Now, I have a little more 
background to give you before I 
get into my subject entitled 
"Personal Computing and the Patent 
System," a subject I happily 
approach from my perspective not 
only as a practicing patent at- 
torney but also as an owner and 
user of a personal computer. 

I am only going to mention 
copyright peripherally because 
another attorney interested in 
personal computing, Ken Widelitz 
from Los Angeles, is giving an 
excellent presentation on that 
subject, and each of us has pro- 
mised not to steal the other's 
thunder, so to speak. So, for 
information on copyright I refer 
you directly to Ken's presenta- 

It is my thesis that the 
encouragement which our Patent 
System has given countless thou- 
sands of inventors has made per- 
sonal computing possible. A little 
bit more history, and then we will 
define patents, explain how they 



BOX 1579, PALO ALTO CA 94302 

are obtained and used and touch upon 
the dilemma confronting us as to 
patent protection of software. 

The Telephone Grandparent 

More than one hundred years 
ago in 1876, Alexander Graham Bell 
invented the telephone and secured 
a patent on it. 3 Since that time, 
the telephone has grown to be an 
absolute necessity of modern 
life and it is not surprising to 
learn that today the telephone 
company, AT&T, has more U.S. Pat- 
ents currently in force than any 
other single entity, approximately 
ten thousand patents, 4 covering 
everything from plastic materials 
used in construction of home 
instruments to fiber optics, 
to patents describing highly 
automated, complex switching equip- 
ment that makes direct distance 
dialing a reality. And, it is this 
latter technology that is in my 
opinion one of the two grandparents 
of personal computing. You see, 
forty years ago, the telephone 
company had an acute need for a 
new theoretical approach to the 
design of automatic long distance 
dialing equipment. It was a bril- 
liant Bell Telephone engineer, C.E. 
Shannon, who, in 1937, and fortu- 
nately for all of us interested in 
computers, recognized the potential 
of the work of the Nineteenth 
Century English philosopher George 

...Boole, and. developed from BopJLe '..s 

works what we now call boolean 
algebra — an analytical tool for 
electrical circuits. 5 (You know: 
boolean algebra is the algebra of 
the binary number system, the base 
two, on-off number system, a 
system understood and utilized by 
our personal computers . ) 

Curiously, or perhaps not so, 
the telephone company sired the 
other grandparent of personal 
computing. While some of the Bell 
Telephone laboratory scientists 
were hard at work applying boolean 
algebra to long distance switching 
problems, three other billiant 
scientists at Bell Labs, Bardeen, 
Brittain and Schockley, were invent- 
ing the transistor in 1946.6 i n a 
book published in the early 1950' s 
entitled Player Piano , author Kurt 
Vonnegut, Jr. futurized about a 
centralized national computer he 
called EPICAC of epic capability, 
installed in Carlesbad Caverns, 
New Mexico. 7 You see, to Vonnegut * s 


prescient mind, it was a vacuum tube 
computer. Had he then known about 
the future of the transistor, 
Vonnegut ' s story might have been an 
even closer projection of what we 
have today. As written, Player 
Piano was an incredible insight 
into the future, including social 
dangers which might flow from the 
misuse of computer power. 

Because of the transistor, 
your personal computer in your home 
or office is potentially more power- 
ful than Vonnegut' s Carlsbad com- 
puter or any vacuum tube computers 
that could ever have been built. 
The transistor and all of its 
progeny are the subject of literally 
thousands of U.S. letters patent, 
granted to thousands of inventors 
who have so rapidly advanced our 
semiconductor technology, from 
point contact transistors, to 
junction transistors, to epitaxial 
transistors, to field effect devices, 
to low scale integration, TTL, medium 
scale, and now to what we probably 
inaccurately refer to as large 
scale integration, inaccurate in 
.that we are now pushing packing 
densities upwards with such patented 
technologies as V-groove^ and high 
resolution masking equipment. It 
probably will not be too long 
before we will see, for example, 
a single chip 64K bit random access 
memory device. Think about it — the 
possibilities that lie ahead. And, 
it has bu&n the disclosure s> in- patents 
that have provided inventors with 
the stepping stones of essential 
technical information which have marked 
the path that has led us to where we 
are today — that brings us here 
together at this Computer Faire. 
Now, hopefully, I have con- 
structed enough of a background or 
operating environment to make the 
subject of patents, if not interest- 
ing, which is my belief and goal, at 
least palatable. And with that, let's 
consider some definitions. 

Inventions and Patents Distinguished 

Let ' s begin by comparing and 
contrasting the terms invention and 
patent . For every patent there must 
be an invention, but it does not 
follow that there will be a patent 
granted for every invention. In 
a broad sense each of us in an in- 
ventor creating an invention when- 
ever, through exercise of our men- 
tality and skills, we devise some- 
thing new. The dictionary defines 

106 BOX 1 579, PALO ALTO CA 94302 

invention as "a device, contrivance, 
or process originated after study 
and experiment." Another defini- 
tion worthy of note accompanies 
the word "invent": "to produce 
something useful for the first time 
through the use of the imagination 
or of ingenious thinking and experi- 
ment, such as a new machine." Even 
in the popular definition of inven- 
tion we find the element of novelty, 
that is, an invention must be some- 
thing new or novel. 

Before a patent may be granted 
under our U.S. patent laws, another 
essential ingredient must be pre- 
sent: "unobviousness," a word of 
art which, while easily explanable, 
is somewhat difficult to apply in 
practice. It is said that a patent 
will not be granted unless the inven- 
tion is one which is not obvious 
to a person having ordinary skill in 
the particular technology to which the 
invention pertains. Let's apply this 
test by an example. 

Suppose the invention is a new 
design for a microcomputer utilizing 
a bidirectional data bus for communi- 
cation between central processing 
and memory. Let's further assume 
that the design is truly novel, that 
is, that no one has ever made a 
microcomputer that looks just like 
this one. Let's also assume that 
the person of ordinary skill is a 
graduate electrical engineer of 
average skill and five years ex- 
perience in the design of electronic 
digital computer systems. Please note 
that my assumption as to the level 
of skill is purely arbitrary and not 
to be relied upon outside this illus- 
tration. * 

Now, let's put this hypotheti- 
cal engineer of ordinary skill in a 
laboratory. On the walls of the lab, 
let's tack up the closest prior art 
references we can find which describe 
similar computer systems. Now comes 
the test: if our mythical engineer 
is able to synthesize our new inven- 
tion from a combination of these ref- 
erences with his ordinary engineering 
skill, then the invention is not 
patentable. It is said to be "obvious" 
or too obvious to merit a patent. On 
the other hand, if the engineer cannot 
synthesize the invention from the 
references, then a patent should be 
obtainable. So we see that the sub- 
ject matter of a patent is a patent - 
able invention. 

Patents are issued to inventors. 
Since patents are intangible personal 
property, they may be transferred, 
licensed to and/or owned by a party 
other than the inventor. This is 
often the case where engineers are 
hired by companies to make developments 
and innovations, and some turn out 
to be patentable — in this situa- 
tion the employer usually owns the 
invention, particularly if there is 
a written employment agreement which 
says so. Where inventions are made 
outside the scope of employment, usu- 
ally the inventor owns the invention, 
although an employer may be able to 
claim a limited "shop right" in the 
invention if its time and/or resources 
were used by the inventor in making 
his or her invention. This is a 
tricky area, and it is not safe to 
suppose that there are clear cut 
rules or results. There are a lot 
of lawsuits filed in this area, 
particularly in the case of so-called 
"spin-offs" where a group of employees 
leave together and set up their own 
competing company. 

The duration of a patent is 17 
years from the date of issuance, 
non-renewable. The average time 
today for the examination of pat- 
ents is approximately 19 months 
from the date of filing to final 
disposition, 9 either issuance of 
a patent, abandonment of the appli- 
cation, or appeal. Currently there 
are approximately 1265 patent ex- 
aminers at the Patent Office in 
Washington, 10 and each has at least 
a technical background, with some 
of them also being trained in the 
law. These Patent Office examiners 
administer the patent laws on a day 
to day basis. An applicant unsatis- 
fied with an examiner's negative 
action has appellate remedies, first 
to the Board of Appeals in the Patent 
Office, and then if still unsatisfied, 
to the Court of Customs and Patent 
Appeals, a five judge federal court 
sitting in Washington, D.C. 

There is only one patent office 
in the United States because it 
is an exclusive activity of the 
federal government. 

States do not issue patents, 
although this has not always been 
true, and Massachusetts was the 
first to grant a patent in colonial 
days. In 1641 one Samuel Winslow 
was granted a ten year exclusive 
right by Massachusetts covering 



BOX 1579, PALO ALTO CA 94302 

his particular process for making 
salt. U.S. Patent No. 1 was issued 
to a Philadelphian on April 10, 
1790 for his apparatus and process 
for making potash and pearl ash. 
U.S. Patent No. 4,000,000 was is- 
sued on December 28, 1976 to a man 
from Las Vegas for a process of 
recycling asphalt-aggregate com- 
positions. Well over 1000 U.S. pat- 
ents are issued every Tuesday, fifty- 
two weeks of the year. In 1976, 80,735 
U.S. patents were issued, and in the 
same year 109,227 new applications 
were filed, which suggests that almost 
three fourths of the total applications 
filed resulted in issued patents. 

A patent application today 
requires a minimum $65.00 filing fee, 
and if issued as a patent, a minimum 
issue fee of $112.00. These fees 
are paid to the Patent Office in 
accordance with federal law. They 
are above and beyond attorney's 
fees and charges for such things 
as making the required drawings. 

Contents of a Patent 

Before I summarize what has 
to go into a patent, let me tell 
you the theory behind patent dis- 
closures. It is like a bargain or 
deal, a contract between the inven- 
tor and the people represented by 
the federal government. The inven- 
tor gives up a full and complete dis- 
closure of his invention — it is no 
longer a secret — with sufficient 
details to enable a person skilled "in 
the particular technology not 
only to make the invention but also 
to make it work and use it for its 
intended purpose. In exchange for 
this disclosure, the govern- 
ment gives the inventor a patent 
for 17 years. 

We will talk about the rights 
that a patent grant conveys in a 
moment, but let's first briefly 
mention the four requisites of a 
patent application: a specifica- 
tion, an oath or declaration of 
the inventor, drawings when neces- 
sary, and the prescribed filing fee, 
which we have already mentioned. 

The specification, which is 
the required written disclosure of 
the invention, has several parts which 
I will list: title, abstract of 
the disclosure, background of the 
invention, a summary of the inven- 
tion, brief description of the draw- 
ings, detailed description of a pre- 
ferred physical embodiment of the in- 

vention, and last, but certainly 
not least, the patent claims, 
which define in typically stilted 
yet very precise language the 
boundaries of the invention, just 
like legal descriptions in land deeds 
describe the boundaries of the real 
property conveyed thereby. 

When an application ripens into 
a patent, there is the grant it- 
self, and the document looks im- 
pressive, with blue ribbons and 
a red seal. 

Enough of theory, let's con- 
sider an exemplary patent, one I 
feel is appropriate for this 
audience. It is U.S. Patent No. 
3,821,715 which issued on June 28, 
1974 and is owned by the Intel 
Corporation. This patent describes 
and claims a general purpose digital 
computer formed out of large 
scale integrated circuit chips: 
one chip being a central proces- 
sing unit (CPU) , the second a 
random access memory (RAM) , and the 
third a read-only memory (ROM) . 
Here now is the front page of the 
specification which includes the 
cities of residence of the three 
co-inventors, Intel as the owner or 
assignee, the prior art references 
cited by the Examiner, an abstract 
summarizing the invention, and one 
of the figures of the drawing, a 
figure which is supposed to be 
most representative of the overall 

United Slates Patent 

HuK, Jr. rt al. 

ini 3,821,715 
l«l June 28, I974 

|M| mfmorv syvit m nm a milti-chip 

|75| Inventory Martian M«ard IMI. Jr., S:.nta 
Clara. Sbfriry Mm*. Sunnv* Jo; 
Fedrrica rasrfa. Cupertino, all ■* 

|T1| Altitun: Inttl ftrponlkKuSunta Clara. Calif. 
|22| Filed: 
|2I| Appl No: J254M 

1 52) US. 0..... 340/172.5,34071 73 R, 34C/I73 SP. 

1ST | lot. CL 00*1 13/00. Gl Ic 1 1/44 

|SS| Field «f Search 340/173 5, 173 SP, 173 R; 

307/238, 279 

1*1 . 

3 jU 1.763 

References Ctod 

2/1972 Cncchictal 

7/1972 AilubHal 

1/1972 Mcxtettal. 

307/2 J«X 
. 340/173 R 
. 340/173 R 


11/1973 Hancyelal _ 

3/1973 Cappos _ 

- 340/172.3 
. 340/173 R 

3.73l.3«« 3/1971 Del .VMII7J % 

3.731.U.I J/1973 HrjuvM VUVI73R 

3.7S7JW*. VIV1J Cum.-. 3«»l7:j 

1,7*1.7 J 1 WIV73 Hcanuhil «l A 3WI12 i 

Schucnrnnnn. "Compuur CiwlroT in IBM Technical 
Drrclntwv- Hulk-tin. Viil 14, No. 12, Xla> 1472; pp. 

-Paul J. lfcnon 
AtshkM EMmntr— Mclvin H. Chapnick 
AUamcj. Agent, or Firm— Spcmlcy. Horn & Lainiu 


A general purpose digital computer which comprises 1 
plurality of mclnl-oxidc-scin'condiictor (MOS) chips. 
Random-acresvmemorics (KAM) and rcad-only- 
memorics (ROM) used as pan of the computer are 
coupled to common N-Jircciional data huscs to a cen- 
tral processing unit (CPU) with each memory includ- 
ing dwJrcf, circuitry to determine »h;ch of the plu- 
rality of memory chips is being addressed by the CPU. 
The computer is fabricated using chips mounted on 
standard 16 pin dual in-line packages allowing addi- 
tional memory chips to be added to the computer. 
17 Cbims, 5 Dra«te„' ~ ■., 



BOX 1579, PALO ALTO CA 94302 

There are five drawings in 
this patent. In a case such as 
this, drawings are required and 
must illustrate every feature of 
the invention. The drawing on 
the front page, above, is actually 
Fig. 1 and is a general block 
diagram of the computer. The 
figure reproduced immediately 
below is Fig. 5, and is said to 
be a detail block diagram of a 
RAM chip. Since this invention 
concerns novel computer system 
architecture rather than circuit 
details, the invention is illus- 
trated entirely in block diagrams. 
Where the invention concerns novel 
circuitry, schematic circuit 
diagrams would probably be re- 
quired. Here now is Fig. 5. 




&&&&/ Amw&as 

I 3 MU.7T/XS-t&? 

From this Fig. 5 drawing, you 
can appreciate that it depicts a 
four parallel bit, dynamic memory 
having a total array size of 256 
bits — a very simple device 
compared to the 4K and 16K devices 
now available or in late product 
development . 

There are two other illustra- , 
tions I want to give you. The next 
is the first page of printed speci- 
fication, and conveys the flavor of 
the language in which patent appli- 
cations are usually written. 
While the language is clear, it 
is technical. It does not read as 
easily or as interestingly as a 
Vonnegut novel, for example. 




1. Field of the Invention 

The invention relates to the field of digital comput- 

2. Prior Art 

Since their inception, digital computer applications in 
have evolved from calculations thrown data press- 
ing and into the area of control. In recent \earv unh 
the development of the so-called "mini computer. " ap- 
plications, particularly in the control area, have ereatly 
increased. Mini computers today arc used at the heart ) 5 
of many systems since they are more flexible, can be 
easily personalized for a particular application, can be 
•ore readily changed or updated than fixed lope de- 
sign systems, and. most s ; znifkantly. the cost ci such 
computers is much less than the cost of a targe general 20 
purpose digital computer. Unfortunately, the size and 
cost of even the smallest mini computers has limited 
their use to relatively large and costly systems. Because 
of this, many smaller systems are fabricated with com- 
plex hardwired logic circuits. 25 

The present invention provides a general purpose 
digital computer which may be fabricated for a cost 
considerably less than the cost of r.en the srr-iilesc 
mint-computers. As will be seen, the presently dis- 
closed computer can provide the same arithmetic con- 3° 
Crol and computing functions as a mini-computer with 
a few MOS chips with these chips being housed in stan- 
dard packages. With the presently disclosed computer. 
it is anticipated that entire new applications, untapped 
because of the cost of mini-computers, will become 35 
practical. These applications include control functions. 
such as numeric controls, elevator control, highway 
and rail traffic control, and process control. The dis- 
closed computer can be used as computer peripheral 
equipment to control displays. ke>noards. printers, ~*° 
readers, plotters and terminals. Other applications for 
the presently disclosed system include computing sys- 
tems and countless other applications within the fields 
of transportation, automotive uses, medical electronics 
and testing systems. 4S 


A general purpose digital computer which compricss 
a plurality of separate MOS chips is described. The 
chips are interconnected by a number of lines, includ- 
ing four bi-directional data bus lines. One chip includes 
a central processing unit that is coupled by the bi- 
directional lines to a plurality of memory chips which 
include random-accevs-memories (RAM) and read- 
only-memories (ROM). The ROMs are used to store 
the computer instructions and other data. A plurality ot 
separate RAMv and. ROMs may be added to -tiie bi- 
directional lines. The memory chips each include de- 
coding circuitry for recognizing a predetermined code. „. 
thus permitting the central processing unit to address 
a single one of the plurality of memory chips even 
though all the memory chips arc coupled to the com- 
mon data bus lines In the presently preferred embodi- 
ment.eachofthechipsare mounted on standard 16pm ,, 
dual in-line packages :ind input and output information 
to the computer is read in thiough and read out from 
terminals on the memory chips. 


FKj. 1 is a general Mock of the disclose.: 
computer. It illustrates the cemrd privessmg unit. * 
angle RAM and a single ROM 

f-tU. 2 is a graph illustrating a **rzle mMruriion cycle 
of the computer and is used pnirarily to describe the 
manner in which the memories cctnmunic^;e with the 
central processing unit. 

FKJ. 3 is a hk«ek digram illustr^nig the inierconncc 
turns between the central processing unit and a plural- 
k> of memory chips including ROMs .:;id K \Ms. 

MG. 4 is a detail block diagrar: illustrative a single 
ROM such as the ROMs illustrate in FIG 3. 

FIG. 5 is a detail Nock diagram ci a single RAM such 
as the RAMs illustrated in FIG. 3. 

Referring first to FIG. 1, a corrsuter bum in accor- 
dance with the present invention a illustrated which in- 
cludes a central processing unit or central processor 
10, a random-access-memory (RAM) 35 and a read- 
only-memory (ROM) 30. In the presently preferrej 
embodiment, four bi-directional cita bus lines 20, 21. 
22 and 23 arc utilized to comrrunicate lt formation 
from the processor 10 to the memories and also tc 
communicate information from -ic memories to the 
processor. Infcrma'.io^ may be re^J Iron the computer 
on lines 56a. 56/j. 56c and 56rf and informaiior: may be 
read into or from the computer on the ir,put;outpu: 
lines 57a, 576. 57c and S7d. As will be explained ra 
greater detail, the central processing unit or processor 
10, the ROM 30 and the RAM 35. are eacn fabricated 
on separate MOS chips utilizing MOS technology anc 
are interconnected hy the various lines illustrated, in- 
cluding the common data bus lines 20, 21, 22 and 23. 
These bus lines may be fabricated on a priced circuit 
board along with lines 62, 63, 64, 65. 33 ar.d 51. and 
then connected to the processor 10 and the memories 
30 and 35 v-here appropriate, ir. ;he prt^ently pre- 
ferred embodiment the processor 10 and the memories 
30 and 35 are packaged in dual in-line 16 pin packages 
as is commonly used in the semiconductor industry 
and, as will be seen, the pins on each parage ha'.e 
been utilized such that the central unit is 
able to communicate with the memories and the entire 
computer is able to communicate througn the port* 
provided on the memories with t-temal circuit rv 

While in FIG. 1 only a single ROM is illustrated, as 
will be explained in detail, numerous ROMs may be uti- 
lized and may be coupled to the common da:a bus linn 
20, 21, 22 and 23. Additionally, a plurality cf RAMv 
may be utilized and these memory would likewise be 
coupled to the corrmon data bus iines 20, 21. 22 and 
23. The central processor 10, as oo the memories, te- 
ceiie complementary timing signals which may be ex- 
ternally generated, on leads 62 and 63 identified as <£, 
and 100 t . respectively, and illustrated :n FKj. 2 as sig- 
nals 66 and 67. Processor 10, in sedition to developing 
other signals, develops * synchronization signal which 
comprises a pa\\c ceneraied every eight periods of the 
signals 4>\ and <fr-. I his lynchror.i/ation is illus- 
trated as signal 6H in FIG. 2 and is communicated from 
the processor 10 M the various RAM* and ROMs uti- 
lised in the computer on k*yd 64. The central processor 
also generates a ROM control signal which is communi- 
cated to the ROMs utilized in the computer via lead 33 



BOX 1579. PALO ALTO CA 94302 



Finally, I have included claim 
1 of the patent below so that you 
see how precisely claims are 
written. The claims are the heart 
of any patent and they are drafted 
to be broad enough to cover the 
invention adequately, but not so 
broad as to be invalid for attempting 
to cover the prior art. Here is 
claim 1: 

We claim: 
35 1. A general purpose digital computer comprising: 

a central processor disposed on a first semiconductor 

a plurality of bidirectional data bus lines; 

at least a separate first and second semiconductor 
4() memory chip each dcfiiilng a memory and each in- 
cluding a chip decoding circuit for recognizing it 
different predetermined code on said bidirectional 
data bus lines and for activating a portion of said 
memory upon receipt of said predetermined code, 
said data bus lines interconnecting said processor 
and said first and second memory chips for com- 
municating said different predetermined codes 
from sai.1 processor to at least one of said first and 
second memory chips and for communicating data 
signals for one of said first and second memory 
chips to said processor; 

whereby said processor may communicate signals to 
said first and second memory chips and said decod- 
, , ing circuits shall determine which memory is being 

If you have taken the time to read 
claim 1, you will note that it is 
one long, single sentence which first 
describes structural relationships 
and ends with the desired functional 
result. The subject of patent 
claims leads directly, I think, 
to the subject of patent rights. 

Patent Protection 

In general concept, a patent 
protects the idea of the invention, 
however that idea may be expressed. 
This protection is broader than 
that obtainable under the copyright 
laws which only provide a remedy 
against copying the particular 
expression of an idea. While pat- 
ents protect inventions and copy- 
rights protect expressions of 
ideas, there is no prohibition 
that in some circumstances a pat- 
ent as well as a copyright could 
be claimed for complementary 
aspects of the same subject matter. 
Usually, however, that which is 
patentable is not copyrightable, 
and conversely so. 

A patent is distinct from a 
trade secret, in that the subject 
matter of the patent is known be- 
caupp it is published when the patent 
is issued, whereas a trade secret 
must truly be kept as a secret to be 
protected. Also, a patent protects 

the inventor against later inventors 
who arrive at the same invention throug 
independent effort. Trade secret pro- 
tection may not be claimed against a 
third party who discovers the secret 
honestly and independently of the 
owner of the trade secret. A good 
example of a trade secret is the 
Coca-Cola syrup formula. Anything 
that can be reverse engineered is 
not a suitable subject for trade 
secret protection. Patent applica- 
tions are kept strictly confidential 
by the Patent Office before issuance, 
and are not disclosed if abandoned. 
Therefore, the subject matter of 
a patent could be protected as a trade 
secret before the patent issues. 

A patent grants to its owner a 
right of exclusion: that is, the 
right to exclude others from making, 
using or selling the invention during 
the life of the patent. This is a 
negative right. It is not the right 
to make, use or sell the invention 
claimed in the patent. Here is an 
illustration: in claim 1 reproduced 
above, the language calls for "a 
central processor disposed on a 
first semiconductor chip." Now, 
it is probable that one cannot 
"dispose" (that means "make") a 
CPU on a semiconductor chip without 
infringing patents covering the 
fabrication of large scale inte- 
grated circuits. So, in order to 
make a single chip CPU, one would have 
to be sure that patents covering chip 
--fabrication were not being inf ringed: - 
or more likely one become licensed 
under such patents to avoid any claims 
of infringement. Another reason 
that the patent grant is a right 
of exclusion is that in some situa- 
tions, other laws or public policy 
prohibit the manufacture or use of the 
invention — examples might be cycla- 
mate artificial sweeteners, and 
refined cocaine. 

The patent grant extends through- 
out the United States and is enforced 
by a suit for patent infringement 
in federal district court. For 
foreign protection, a patent must 
be obtained or confirmed in each 
country, which can be an expensive 
proposition. I cannot possibly 
cover the subject of foreign patents 
in the time allotted, other than simply 
to mention that whole area as a 
separate problem that has to be 



BOX 1579, PALO ALTO CA 94302 

Time for Filing Application 

A patent application may be 
applied for at any time, subject 
to some highly technical, but 
important exceptions. The appli- 
cation must be filed within one 
year of being patented or described 
in a publication anywhere through- 
out the world, and within one 
year of first being publicly used 
or on sale in the United States. 
For inventions made in the United 
States, foreign applications may 
not be filed without a U.S. Govern- 
ment license or awaiting six months 
after first filing a U.S. Patent 
application. Usually, it is wisest 
to file a patent application as soon 
as the invention is completed. An 
invention is deemed complete when 
it is either physically made and used 
— called an actual reduction to 
practice, or a patent application 
is filed with a full disclosure of 
how to make and use the invention — 
called a constructive reduction to 

I cannot overemphasize the need 
to keep adequate notes of inventive 
ideas when the invention is first 
conceived. A bound notebook is ideal, 
and the entry ought to include a 
witness* name and date of reading 
and understanding the invention. 
Sometimes it is crucial to have such 
records when the same invention is 
made by two independent workers, 
and it becomes necessary to prove 
who was first. In any event, a 
policy of keeping an invention note- 
book is a must for any serious inven- 

Patent Searching 

Before an inventor files a 
patent application it is usually 
wise to make a patentability 
search to see if there are prior 
patents or other references which 
disclose or clearly suggest the 
invention. Most patent searches 
are made in the public search 
library maintained by the U.S. 
Patent Office in the suburbs of 
Washington, D.C. A number of public 
libraries around the country main- 
tain collections of U.S. patents 
as well as some classification 
information. In this geographical 
area, the Sunnyvale public library 
maintains copies of U.S. patents 
issued since about 1960. It is 
possible for inventors to make 

searches themselves in facilities 
like the Sunnyvale Patent Library, 
but the work is arduous, and the 
results are generally far less 
reliable than if the search were 
maae by professional searcher in 
the Patent Office search room. 
Patent attorneys maintain close 
ties with searchers in Washington, 
and patent attorneys may be located 
by looking in the yellow pages 
under "Patent Lawyers" which is 
a separate classification from 
attorneys in general. 

Patentable Subject Matter 

Perhaps you have noticed that 
I have not yet told you what is 
patentable and what is not. This 
has been intentional. I have 
deferred that topic until now 
because it leads us right into 
computer programs and patentability, 
my last topic. 

Patents are supposed to promote 
the useful arts. Therefore, patent- 
able subject matter must be useful, 
that is, capable of being used for 
some purpose or function. An 
abstract idea, such as a machine 
for generating the sound of one 
hand clapping, would not be useful. 
So also, perpetual motion machines 
are said not to meet the utility 
requirement. Anything that simply 
will not work, that purports to 
defy the laws of physics or chemistry 
may run afoul of the utility 
requirement . 

Assuming that the subject 
matter is new and useful, it must 
fall within one of the four cate- 
gories of process, machine, article 
of manufacture or composition of 
matter . 

Process means method and pat- 
ents for processes may be concerned 
with methods for making chemicals, 
forming articles, and making measure- 
ments or calculations with electrical 

The difference between a 
machine and an article of manufacture 
is that a machine is said to have an 
inherent law of internal operation 
whereas an article does not. 
Another claimed distinction is that 
a machine has moving parts whereas 
an article does not, but this 
second criterion breaks down with 
personal computers which have 
sophisticated rules of internal 
operation but no moving parts 



BOX 1579. PALO ALTO CA 94302 

(except perhaps the cooling fan 
for the power transformer) . 

A composition includes mix- 
tures of matter and also compounds. 
Chemical patent applications are 
subject to special rules and are 
drafted a little differently than 
are mechanical and electrical 
applications. There is a great 
deal of overlap in such areas of 
electrochemistry as the process 
of fabricating large scale integrated 
circuits on monolithic silicon 

So the four classes of patent- 
able subject matter are process, 
machine, article and compositions 
of matter. Here are some subjects 
that have been held not to be 
patentable: mere arrangement of 
printed matter, articles naturally 
occurring which are unaltered, 
methods of doing business and 
accounting; and, of most import- 
ance to us, the last one is scien- 
tific principles, including 
mathematics and algorithms, which 
brings me directly and inescapably 
to the patentability of computer 
programs . 

Patentability of Computer Programs 

Pure mathematics has tradi- 
tionally been characterized as a 
part of the liberal arts and 
philosophy. It has been held by 
our U.S. Supreme Court 11 that a 
pure mathematical formula or al- 
gorithm is not patentable because 
a patent would wholly pre-empt the 
mathematical formula. In practical 
effect the patent would cover the 
algorithm itself, and deprive all 
but the patent owner of the right 
to apply the algorithm to any 
problem in any technology. In the 
case to which I refer the algorithm 
was a pure algorithm; it had to 
do with the conversion of binary 
coded decimal numbers into pure 
binary numbers. Its application 
was said to be entirely abstract, 
it operated only upon numbers per 
se and was not applied to any 
physical environment or phenomenon. 
This case, Gottschalk v. Benson , 
decided in 1972, has been interpreted 
by lower courts as well as commentators 
as being a very limited decision — 
which is that computer programs 
involving algorithms which do nothing 
more than process numbers without 
any relationship to or impact 
upon the physical world, are not 

patentable. On the other hand, 
the Patent Office has interpreted 
this case much more broadly. The 
present Patent Office position 
appears to be that if the formula or 
algorithm is the only thing new or 
novel, then what is involved is a 
non-patentable mental step, even 
though the algorithm is claimed 
to operate upon something in the 
physical universe. The analogy 
that the Patent Office used in 
recent papers filed with the 
United States Supreme Court 12 
in urging it to review a lower 
court's decision was as follows: 
"It is as if respondent 
[patent applicant] were 
trying to patent the Pythagorean 
Theorem as a part of a method, 
by adding a final step that 
suggested applying the solution 
of the equation to surveying. 
Such a last step is not an 
integral part of a new inven- 
tion, but rather the non- 
inventive application of a 
mathematical result to existing 
technology; that fact is not 
changed by the facile device 
of so drafting a claim that 
it places an abstract mathe- 
matical algorithm within a 
recitation of steps performing 
conventional , existing technology . " 
So let me summarize my personal 
opinion as to where we appear to be 
at the time of the writing of this 
paper (early January 1978) : (1) if 
the only thing that is novel and 
unobvious is an abstract algorithm 
that is not applicable to the physi- 
cal world — no patent. (2) If the 
novel algorithm is applied to the 
physical world by operating within a 
standard general purpose computer, the 
Patent Office will likely hold that 
it is unpatentable, but the Court of 
Customs and Patent Appeals might re- 
verse and grant the patent. (3) If 
the novel software operates in con- 
junction with novel hardware, then 
the Patent Office may then grant a 
patent, assuming other conditions 
for patentability are met. 

One argument that is heard every 
so often is that a unique computer 
program defines a unique computer 
structure when loaded into the compu- 
ter 's program memory. And, literally 
this is true. However, if this were 
the case, then the program which con- 
verted BCD to pure binary would have 
been held patentable (which it was 



BOX 1579, PALO ALTO CA 94302 

not in the Benson case) because the 
computer which that program defined 
would have been unique. So, it 
seems safer to look to physical 
contact, manipulation or impact as 
an indicator of patentability for 
computer programs, at least at this 

From my experience and observa- 
tions, it is my opinion that the 
chances of obtaining a patent involv- 
ing a program are enhanced if one pur- 
sures a strategy of emphasizing the 
hardware and physical contact 
aspects of the invention, if pos- 
sible. For example, if the pro- 
gram resides in the read only memory 
as firmware, then one might claim 
a read only memory arranged to define 
the novel algorithm — this way you 
claim physical structure, not a 
mental step, and hopefully maybe 
you can avoid a rejection by the 
examiner that the claim is non- 
patentable subject matter. 

One cannot mention protection 
of software without noting that it 
is a subject matter which is copy- 
rightable. Whether the copyright 
law provides adequate protection is 
a matter of great national debate. 13 
Remember, we have already noted that 
copyright only protects against copy- 
ing of the expression of an idea, not 
the idea itself. For more informa- 
tion regarding copyrightability of 
software programs, I refer you again 
to Ken Widelitz ' presentation. 14 


In presenting this paper I 
am reminded of a definition I once 
heard for an expert. It has been 
said that an expert is like the 
bottom half of a double boiler — 
it puts out all the strain, but it 
really does not know what's cooking. 
Those of us who work with patents 
sometimes find ourselves generating 
a lot of hot air without really 
knowing whether those warm air currents 
have any utility. I hope you have 
profited from this presentation. 

Today we have discussed the 
need for our patent and copyright 
laws to foster and promote the pro- 
gress of science and useful arts by 
protecting authors and inventors in 
giving them exclusive rights to 
their works for limited periods of 
time. We have noted that for inven- 
tions to be patentable they must be not 
only novel, but also unobvious to the 
ordinarily skilled worker in the 


technology. We have noted that a 
patent requires the* inventor to 
exchange a full disclosure of the 
invention in exchange. for the 17 
year exclusive right to exclude 
others from making, selling or using 
the invention, as claimed. We have 
illustrated the disclosure require- 
ment by looking at parts of Intel's 
patent for a four bit bidirectional 
data bus microcomputer. And, finally, 
we b-^ve noted the problems confronting 
us in the troublesome area of the 
patentability of software. 

It has been my experience that 
electrical engineers and technicians 
tend to be skeptical about patents 
for such things as computers. I 
often hear, well, you can't patent 
Ohm's law, and that, of course, is 
true. But the development of ele- 
gant unobvious hardware and hardware- 
software combinations may be appro- 
priate subjects for patents. Here- 
tofore, the personal computer industry 
as a fledgling industry has not been 
faced with the need to deal with 
patents and the patent system. With 
the popular press projecting annual 
sales in the personal computing 
industry to reach 1.5 billion dol- 
lars by 198 5, I 5 it is apparent that 
this industry will come into contact 
with patents on a more frequent 

Let me leave you with some 
figures that I think prove my point. 
Intel, for example, has about 30 com- 
puter related patents, Hewlett- 
Packard has about 200 patents on 
computer/calculators, and the IBM 
Corporation has over 9,000 U.S. 
patents now in force and effect, 
covering a broad spectrum of sub- 
jects from typewriter ink composi- 
tions to large scale computer 
architectures .16 As a company grows 
in size, apparently so does its 
patent portfolio. 

We have just begun to glimpse 
the possibilities for personal com- 
puting. But what I am waiting for 
and would love to see invented, is 
a truly personal computer , one that 
interfaced directly with my brain. I 7 
Then, theoretically, I could perform 
complex mathematics, compose symphonies; 
translate languages and hopefully 
advance the state of humanity. This 
is one of my hopes and speculations 
for the future of personal computing 
and it is people like you that possibly 
will come up with an invention to bring 
it into reality. Thank you for your 

113 BOX 1579, PALO ALTO CA 94302 

Foot >es 

-The Copyright Laws, Title 
17, United States Code; The Patent 
Laws, Title 35, United States Code 

2 Eureka! An Illustrated History 
of Inventions from the Wheel to the 
Computer , Holt, Rinehart & Winston 
(c) 1974, pp. 24, 39. 

3U.S. Pat. No. 174,465, 
issued March 7, 1876 to A.G. 

information kindly supplied by 
the Patent Department of AT&T. 

5 Burroughs Corporation, 
Digital Computer Principles , 2d. Ed. 
McGraw Hill, ^ 1969, page 94. 

Eureka 1 An Illustrated History 
of Inventions from the Wheel to the 
Computer , supra Fn 2, p. 231. 

TkT Vonnegut, Jr., Player P iano, 

Dell Ed., (c) 1952, pageTf. 

8See, e.g., U.S. Patent No. 
3,924,265, issued December 2, 1975 
to Rogers for "Low Capacitance V- 
Groove MOS NOR Gate and Method of 
Manufacture . " 

9 United States Patent Office, 
Office of Information Services. 

ll Gottschalk, Acting Commis - 
sioner of Patents v. Ben son, 409 U.S. 
63 (1972). 

barker, Acting Commissioner 
of Patents v. Flook . No. 77-642, 
filed Nov. 2, 1977, Petition for 
Certiorari from In re Floo k, 559 F.2d 
21 (CCPA 1977). 

13The Report of the Software Sub- 
committee to the National Commission 
on New Technological Uses of Copyrighted 
Works, issued in June, 1977, opines 
that copyright is the most appropriate 
form of protection for software pro- 
grams and proposes changes to Section 
117 of the New Copyright Act. A 
dissent by John Hersey argued that 
since a computer program was only 
operable as part of a computing machine 
it did not belong in the category of 
copyrightable subject matter and 
proposed a new special law for the 
protection of software, a so-called 
Computer Software Protection Act, sug- 
gesting a 10 year period of protection 
for software. 

14 See Ken Widelitz * Legal/Busi- 
ness Forum column in the November 1977 
issue of KILOBAUD , Page 14. For an 
excellent legal discussion on the 
overall subject, see "Computer Program- 
ming and Patent Law," American Patent 
Law Association Quarterly Journal , Vol. 
V., No. 1, 1977. 

!5 u.s. News & World Report , Nov. 
21, 1977, page 77. 

16information kindly supplied by 
Patent departments of Intel, Hewlett- 
Packard and IBM, respectively. 

17prof. Carl Sagan suggests this 
in his currently popular book The Dra - 
gons of Eden, Speculations on the O ri- 
gins of Human Intelligence , Random 
House, 1977, page 205. 



BOX 1579, PALO ALTO CA 94302 

uupyi lyni. dnG juitwafe; 

Some Philosophical and Practical Considerations 
by Kenneth S. Uidelitz, Attorney at Lav; 
10960 kilshire Boulevard, Suite 1504, Los Angeles, CA (2:13) 477-3067 

At the present time the National Commission on 
New Technological Uses of Copyrighted Works (CONTU) 
is working on a report which will recommend the Man- 
ner in which a Copyright Law should be modified in 
order to protect computer programs. The committee's 
latest report, issued in June, 1977 is available 
from CONTU, Washington, D. C. 2055b, telephone 
(£02) 557-0996. 

The latest CONTU report discusses the three 
currently available vehicles for protection of com- 
puter programs, namely, copyright, patent and trade 
secrecy. As the report states, "...copyright is 
designed to protect the expression of ideas while 
patent's purpose is to protect what are generally 
understood to be inventions -- in a sense the ideas 
themselves." Patents, on the other hand, protect in- 
ventions which must be useful, novel and not obvious 
with those familiar with the related technology. 
Trade secrecy involves the dissemination of informa- 
tion pursuant to contractual agreements by the terms 
of which the party receiving the information pro- 
mises not to reveal it to anyone else. 

Philosophically speaking, issues raised in dis- 
cussing whether software should be protected by co- 
pyright involvea the use of analogies in which the 
statement "computer programs or more or less like..." 
are made. The question becomes, is a computer pro- 
gram a writing and/or a mechanical 'device. Other 
"philosophical" issues are when is a copy of a com- 
puter program made? Uhen it is loaded into memory 
or when it is dumped onto tape or haru copy. Another 
issue involved is when is a program based on a pre- 
vious program a derivative work. That is, an author 
of a book written in English has the exclusive right 
to translate it into french (read BASIC for English, 
COBOL for french?) 

The Copyright Act of 1970 

The Copyright Act of 197C became effective Jan- 
uary 1, 1973. Section 117 of that Act specifically 
stated that it did not change the rights of an au- 

thor of a computer program as such rights 
existed under the old law. CONTU, in its 
report, recommendsa new Section 117. how- 
ever, many of the new provisions of the 
Copyright Act do currently effect computer 

The new Copyright law makes the term 
of copyriglit equal to the life of the author 
plus fifty years. The new law also makes 
the concept of "publication" less important. 
The new law also defines when a work is 
"created", when a "copy" is made and what 
constitutes "a work made for hire." 

Under the Copyright Act an author has 
the right to do any of the following: (1) 
to reproduce the copyrighted work in copies; 
(2) to prepare deriva tig works based upon the 
copyrighted works; (3) to distribute copies 
of the copyrighted work to the public by 
sale or other transfer of ownership, or by 
rental, lease or lending. Each of these 
rights are independent of one another so 
that any one may be retained while any other 
is sold or transferred. 

The question as to what constitutes a 
derivative work is one of the toughest that 
computer programmers will face. It must 
first be understood that copyright protec- 
tion does not extend to any idea, procedure, 
process, method of operation or concept. 
Copyright protection extends only to the ex- 
pression of ideas, not the ideas themselves. 
Cf course, some ideas, concepts, procedures 
or processes are so basically fundamental 
that copyright does not protect an explana- 
tion cf them. For example, some sorting 
routines are so basic that they cannot be 
copyrighted. However, if a series of basic 
routines are put together so as to accom- 
plish a specific task, that program is 
subject to copyright. 

In order to obtain copyright prctec- 

copyright 1977 Kenneth S. Widelitz 

tion, there are certain procedural technicalities. 
These are notice, deposit and registration, no- 
tice requirements are simple. They consist of a 
C in a circle, the words copyright or the abbrevi- 
ation copr. ; the year of the first publication and 
the name of the copyright owner (i.e., Copyright, 
1978, by Kenneth S. Widelitz.) rThere are rules 
relating to where the notice must appear, depend- 
ing upon the type of work to be copyrighted. The 
Copyright law also requires that materials which 
are copyrighted be deposited with the Library of 
Congress. There are exceptions if the Library of 
Congress does not desire specific material. Re- 
gistration with the Registrar of Copyrights is a 
prerequisite to the bringing of a law suit for 
the infringement of copyright. 

Another consideration is the Doctrine of Fair 
Use. Fair use embodies that notion that a rea- 
sonable portion of a copyrighted work may be re- 
produced without the permission of an author for 
a legitimate purpose. The doctrine is most often 
applied to teachers who have xeroxed materials 
for distribution to students for educational pur- 



Ted Lewis 

Every author of a successful book knows the essentia] ingredients of a 
"best seller", whether it be a "great American novel" or a book about computers. 

First, the author must have something to say. This concept is often over- 
looked by an anxious novice eager to get his/her words in print. Yet, unless 
there is value in your words, the book or article will be useless. In computing 
there are many ideas and concepts that need to be clearly and concisely stated. 
Unfortunately, books on computing tend to be re-hashes of the same old concepts 
and ideas inherited from the first 25 years of computing. The microprocessor 
revolution has changed many of the ideas and motivations in computing. It 5s 
this change that may prompt you to write. If" so, be sure to take a fresh and 
innovative approach. 

The second essential ingredient to good writing is style. Clearly vrritter. 
books are as valuable as clearly coded programs. Everyone has a unique style, 
but as authors we owe it to our readers to suppress nonsensical idiosyncrasies 
and "in" human. Perhaps one of the most difficult tasks for an author of 
computer books is to avoid use of mnemonics; clever "buzzwords", and trite 

Style also contributes to a book by forcing a structure on the writing. 
Chapters are the modules of a book; sections are subprocedures , and paragraphs 
are equivalent to control structures. Thus, a good book has a clean structure. 

Style also means the book is organized into a sequence of increasingly 
complex topics. Initially, a ground-level introduction is used to familiarize 
the reader with important concepts used later. Don't include material that is 
minor. Do include material that will help the reader understand later concepts. 


A final suggested ingredient in good writing is timing. A "good" book is 
not always a "good selling" book. The reason is timing. Computer books have a 
"life" of roughly 3 years. Furthermore, it may take 2 years to produce the book 
in the first place. This means the author must look into the future to see what 
will be viable and may be of interest to the future computing community before 
starting a new book. An author must be a prophet of things to come. 

Predicting the future of personal computing is almost impossible. Therefore, 
writing is a risky business. For example, two years ago a wise writer should 
have expected a decline in hit building, and a rise in turnkey business systems 
and disk files. Today, there are several trends leading to "successful" topics 
for authors of 1980 books. 

The foregoing suggests three fundamental concepts of writing a good book. 
There are many other aspects of writing not. covered here. p or example, the 
author-publisher relationship, how to locate a publisher, how to promote an idea, 
etc. All of these factors contribute to the success (or failure) of a writer. 

Nonetheless, the place to start a writing career is with 1) something of 
value to say 7 ~ 2) a polished style, and 3} a topic 2 years ahead of itself. 




Douglas J. Mecham 
Hughes Aircraft Company, P. O. Box 3310, Fullerton, California 92634 


Programmers write neat programs that they 
share with others; thus, there is a need to write 
useful user guides. From the user's standpoint he 
needs an easy and simple guideline to be success- 
ful. While this documentation task may seem 
difficult; the programmer writing the user's docu- 
mentation for his program can also be a success 
easily and simply. The easy and simple approach 
is to learn a few easy thought processes. This 
presentation deals with eight major considerations 
in writing the user's guide. 

Your use of microcomputers has revolu- 
tionized the world. No longer is a highly trained 
specialist required to use a computer. By virtue 
of your attendance at the Computer Faire you are 
sharing with others your experiences. These facts 
require you to communicate your ideas. How well 
do you do that? If you have to describe in writing 
how to use your neat program, how well do you 
communicate? The best idea in the world is not 
worth much unless it is communicated and used; 
in other words, your knowledge is worthless if it 
is not communicated; furthermore, your intelli- 
gence of that knowledge is not credited unless your 
ideas are used effectively. 

The reason we should be concerned about 
the details of documentation is because we need to 
communicate information in writing about our 
ideas. This is important now because this is the 
decade of the user. More people, non-computer 
oriented people, have direct access to computer 
systems. These users do not tolerate the way the 
system wants to operate but want to dictate to the 
system how they wish it to operate. Good user 
documentation is the key to effective use of a 
computer system. This paper will provide some 
useful ideas to generate good user documentation. 

To provide perspective of where user 
documentation fits consider the three parts of a 
computer system: 



It is within the category of Peopleware that docu- 
mentation falls. The objective of the user 
documentation is to break down the barrier between 
the real user and the computer system hardware 
and software. 

Written information is a personal thing 
specifically designed for particular situations; 
however, there are several aspects of it that can be 
identified. This paper discusses these aspects; in 
summary they are: 

WHO- -understand the user's thought process 
SIMPLICITY— keep it simple 
FORGIVENESS- -make recovery easy 
EXAMPLES --use plenty of meaningful 

DETAILS --sink the details 
RETRIEVAL- -provide easy access 
PERSPECTIVE— keep the user point of 

view at all times 
APPEARANCE --it must look nice to read 
Remember the primary objective of the user infor- 
mation is to make the user successful, easily and 
simply within his time/dollar constraints. End 
users are very impatient about time and money. 


Consider who the user /reader is. The 
usual frame of reference for a writer is his own 
thoughts but for a user document the frame of 
reference must be the user. You need to determine 
the user's psychological profile and understand his 
thought processes. This is not so difficult since 
you are well aware of how you think as a user of 
other peoples' software. The type of information 
presentation a user expects when reading a user 
document can be determined by choice of form, 
format, and vocabulary. 

First, look at what information the user 
needs to perform his task. The capabilities of the 
software should easily define the parameters in- 
volved. Not all the information may be required 
for all users so the user document must present 
different amounts of information to meet each 
individual's requirement to complete his task 
successfully. Certainly there is a minimum amount 
of information necessary; this of course is the 
simple path. Then there is an average amount of 
information required; this would be the normal path 
a user would follow. There are always those 
complex tasks that require the user to need the 
maximum amount of information. Aside from these 
different paths, the most necessary set of user 
information needed is when an error or problem 



BOX 1579. PALO ALTO CA 94302 

Second, once the different sets of informa- 
tion have been determined they must be ordered in 
some manner. This implies a sequence of events. 
Again, this determination is a function of what the 
user needs; that is, the sequence of events as the 
user perceives them. An easy way to determine 
the sequence of events is to ask yourself "What is 
the first element of information the user needs, " 
"what is the second element of information the user 
needs, " and so on. Do not fall into the trap of 
providing too much information on the account of 
"what if;" consider only the critical information 
necessary. The user does not want to wade through 
superfluous information. The user document is not 
a design specification document and as such the 
flow of information in the document is by user task, 
not by technical item commonality or logic. 

One of the user viewpoints often left out is 
describing what the user cannot do. Remember the 
user's perception of what he expects the software 
to do may not coincide with what the software can 
do. The user most likely will attempt to do what 
he thinks he can do with the software, or what he 
thinks he should be able to do with the software. 
Assume very little and leave little to the imagina- 
tion of the user. 

How ever you plan your approach to meet 
the above criteria you need to plan what information 
is needed, the information flow, and the information 
grouping. There is nothing that says that all 
information groups must be mutually exclusive; do 
not be afraid to replicate elements of information 
within several groups. The sequence is important 
to the user. The information should be in such a 
sequence to accommodate the novice or student 
(simple), the experienced user (average), and the 
computer systems nut (complex). 

Notice that the real information needs are 
based on what the user wa nt s, not on what you think 
he needs. It must be kept in mind that the user 
requires only that information necessary to get him 
to the next step to complete his task. An incredible 
erroneous concept in data processing is that user 
documents should define all their technical terms 
in the front of the document. Then. . . they expect 
the reader to remember and understand them for 
the duration. If you need to define a particular 
word then define it when you first use it in a mean- 
ingful context. Do not be afraid to redefine it later; 
most technical readers have a short memory so 
why make it difficult for them. 

For any good user document the first 
paragraph should indicate: 

Who should use the document, 

Why he would want to use the document, and 

How to use the document (conventions and 
This way the user may readily find the information 
he requires without going through useless material. 
The user document is usually a reference type 
document, not a novel. 

Vocabulary is a sore point with data 
processing persons, or for that matter persons in 
any other technical field. The reason for this is 
because "it is obvious what a term means. " Un- 
fortunately, the reader cannot always rely on 
context to figure out the meaning of a word. Con- 
sider the home computer user or any other non- 
data processing professional coming in direct 
contact with the computer. The words chosen must 
not let him think he is getting "computerese" all 
over him. While some relish the acronym, jargon 
is always hard to keep straight. Why not use words 
that are easily recognizable by the user. Using a 
meta -language (a specially -defined translation 
language) is a NO-NO- -spell it out! 

Take, for example, the error message 
ILLEGAL or FAILURE and consider the psycholo- 
gical ramifications on a law-abiding successful 
business man. The psychological impact of a user 
touching the computer is great ! At first the user is 
afraid he will break something or "it" (the com- 
puter) will "do something" that will render htm less 
capable to carry on in his life. Or "it" will des- 
troy everything he has worked hard to achieve. 
The user may think the computer will violate his 
sense of goodness. One purpose of the user's 
document is to put the user at ease and dispel any 
such fears. Thus, the user's anthropomorphic 
sensibilities are alleviated. 


The basic rule is KEEP IT SIMPLE ! 
Whether the user is a sophomore or holds a PhD 
his desire, or his ability, to read is small. All 
the user wants to do is get his task completed, 
simply and easily. Simplicity may be used at all 
levels of user documentation. Technical infor- 
mation tends to be complex, whether it ts or not. 
All too often numerous concepts and parameters 
are juggled before the user's eyes. A psychologist 
once told me that a person can generally only 
handle eight major items at once. 

A user may understand sophisticated 
vocabulary and complex concepts. But, each time 
a user's thought process must make a translation 
or complex transition the probability for loosing 
information or concepts is significantly high. 
What the user does not need is confusion. A sim- 
ple example of this are references to octal or hexi- 
decimal values instead of decimal values. As you 
will see, simplicity also takes into consideration 
who the user is. 

One technique to keep a user document 
simple is to take a "storybook" approach. That is, 
structure the information in a simple, straight- 
forward, step-by-step manner. When organizing 
the information choose a critical path for the user 
to get his task completed whether it is simple, 
average, or complex level. In order to meet the 
"storybook" criterion sentences need to be simple. 



BOX 1579. PALO ALTO CA 94302 

The "fog" factor of long sentences and large words 
requires education beyond most users and consi- 
derable translation. Remember that television is 
geared for the ten year old and it is a very 
successful communications media format. 

If you should leave out information by 
assuming that the user should know it, chances are 
you will introduce confusion. Why make the user 
think more than he has to; certainly user documen- 
tation is not an examination. Why make it difficult 
for even the most sophisticated user. 

Vocabulary plays a role in simplicity. To 
keep it simple use common words known to most 
readers. . .like the English language. Computerese 
and FORTRAN do not have commonly known voca- 
bularies. If you do use common words chances are 
you will relate to your reader and he will under- 
stand your concept. For instance, to explain data 
base concepts use school information for an 
example rather than nuclear power; most people can 
relate to information about schools. While there 
may be some particular vocabulary words required, 
minimize their number. Why should a user learn a 
whole new vocabulary just to do his simple task? 

Another method of simplicity is to make the 
wording tight (no extra verbage) but not terse. 
Leave out computer -oriented words. The five cent 
and ten cent words put together well are worth 
more than a fifty cent word that does not fit. When 
you choose the use of a particular word, be 
consistent and don't use'another word that means 
the same thing or almost the same thing. This is 
especially true when defining new terminology; 
leave the word variety to the novelists. Consis- 
tency makes simplicity. 


Naturally, if there is a way a mistake can be 
made the user will find it. This is easily done 
because the user performs his task the way his 
memory tells him to, not the way the computer 
memory wants him to. 

When there is a conflict the computer sys- 
tem should adapt to the user but this is usually not 
the case. Therefore, it is the user's manual that 
must make the computer system potable to the 
user's way of thinking, in the user's terms. Hope- 
fully the computer system will forgive the user and 
give him another chance. The user manual is for- 
giving by discussing alternatives should a problem 
surface, and by not leaving the user to guess what he 
should do. Computer software usually relies on 
the user manual to bail the user out of his problem 
and put him on the right road to performing his 
task. For example, just think of all the error 
numbers you have seen as a result of a problem. 
The user manual must define each error number 
by a clear indication of what happened, what the 
current situation is, and what to do about it. 
Both the computer system and user's 
document must allow for quick and easy recovery. 
Normally, you will waste more than 50% of your 
time guessing and testing your hypothesis as to 

why a problem occurred but a good user's docu- 
ment can often cut this effort in half. The 
vocabulary of the error messages must match the 
vocabulary of the user document discussion. 
Abbreviations and computerese are unwarranted. 

Be kind to the user for he has problems. 
The user cannot and does not want to distinguish 
between errors (specifications violated), need for 
document clarification, or design change require- 
ments. The user only wants to get his task done 
simply and easily. Unfortunately most vendors ask 
the user to make this judgement when he has a 


The first item the user looks for in a user 
document is an example. More specifically, an 
example that most nearly matches his task. Even 
when using the manual as a reference an example 
will often be needed. Thus plenty of examples in the 
user's document are most helpful. 

Since the user is going to reference the 
examples so often, several important characteristics 
should be noted. First, the sequence of events 
within an example must be very clear and you must 
distinguish between what the user does and what the 
computer /software does. To make the example 
understood provide an explanation of the sequence 
of events or results right along side. Since an 
explanation does not leave the reader guessing; 
be explicit, you know what you mean, so tell the 

Examples should be simple, easy to follow, 
and consistent throughout the user document. 
Often times, considerable continuity may be 
achieved by creating several examples around the 
same topic matter. There is nothing wrong with 
duplicating examples or portions of examples. It 
is even desirable to build on a specific example 
duplicating a previous example to illustrate a 
sequence of events. This way the user can relate 
the different examples. Examples are made simple 
if the subjects are easily related to. For 
instance, if file records were discussed, the use 
of bank book accounting may be helpful since the 
reader can identify with it. 

There is no place for clever, funny, com- 
plex, or wierd examples. These types only cause 
confusion. Good examples represent the manner 
in which most users would use the system. 
Additionally, the examples described must have 
been tested and work. 

Finally, be careful of the symbology used in 
examples so the reader can distinguish between 
form, format, and literal requirements. For 
instance, italics are often used for a "type" of 
input such as name whereas upper case may in- 
dicate literal information as in the case of ZEP=. 
Specifying^ carriage return is difficult so a symbol 
such as (CR) might be used since there is no sym- 
bolic equivalent. Be careful of specifying literals/ 
strings in quote marks, not all systems expect the 
user to enter the quote marks with a string value. 



BOX 1579, PALO ALTO CA 94302 

For example, consider the message ENTER "YES" 
or "NO"? Should the user enter the quote marks 
too? If not, then leave them out. 

The more examples you put in your user 
document the better. 


The rule for detail description is to "sink" 
it. In other words, do not get deeply involved in 
beginning paragraphs of a user's document, save 
in-depth discussion for last. Usually the user 
needs only the straightforward simple information 
to perform his task. Included in "details" are 
complex and odd situations which are needed in a 
small percentage of tasks. If too much detail is 
required to perform the task, perhaps software 
redesign should be considered. 

In a user's document confront the user first 
with the easiest and most useful alternatives. Su^;h 
an approach will minimize room for user problems 
to develop. Once the user is a success with the 
simple approach he is ready for more detailed 
material. Be careful to mention how to use an 
item as well as just a description of the item. 
Bulky material such as tables or long lists should 
not clutter up the text. Since such information is 
rarely read from start to finish and used normally 
for reference, move it to an appendix. Besides, 
the appendix is easier to update than the middle of 
a document. 

How much detail should be put into a user 
document? Rules of thumb are: enough to assist 
the user in accomplishing "most" of his tasks and 
enough to solve "most" of the user problems. The 
detail material is where the user limitations, pit- 
falls, and idiosyncracies are discussed. The 
detaLLsare needed Qnly__after_ the ; user has become 
a success easily and simply at least once. 


If information is useful the user must have 
easy access to it. The user document is primarily 
used for reference. Thus, when a user has an 
information need he thinks of an access point and 
looks for that point in the user's document. Again, 
what the user thinks may be different from your 
point of view so consider several access points to 
an element of information. 

The user most likely will choose a word or 
words related to the task he is doing. The user 
will not likely select a word related to design or 
organization of the software. The implication is 
not only to index the user document information 
by user task but to organize the document infor- 
mation by user task. 

An alphabetical index is mandator v in a 
user document of any size. Often a multiple level 
index is helpful so a user can reference a particular 
word as it is related in different contexts. Per- 
mutting words assists to cross reference material 
where the index item is more than one word. 
Likewise, there may be instances where pointers 

("see Also") to other information may be useful to 
a user tracking down information. Vocabulary 
plays an important role in retrieval since not all 
users and authors think of the same term to des- 
cribe a function or other item of information; 
thesaurus words are very helpful. 

When considering the structure of a user 
document, simple sentences, paragraphs, and 
item lists are easy to access as opposed to long 
paragraphical information. Furthermore, the lay- 
out can be extremely helpful for physically retrie- 
ving information. Some of the latter techniques 
include shading of key words or syntax forms, 
marking paragraphs with bullets/dots, drawing 
rectangles around important information, or using 
separate fonts such as italics or script as well as 
bold face type, underlining, simple indentations, 
and titles. 

Tabulation techniques can take the form of 
colored tabs with printing on them; such tab pages 
can have summary user information on them for 
easy access/reference. Simple inexpensive tabu- 
lation may be achieved by putting black marks at the 
edge of a page. Then when the edge of the docu- 
ment is titled the appropriate section is found using 
an index page in the front of the manual . 

The need for easy and quick retrieval of 
information in a user's document is mandatory 
since all technical materia] is referenced after 
being read. Not being able to easily and quickly 
find information accounts for a significant amount 
of a user's wasted time and frustration. The 
attitude "it's all there. . . somewhere" is not helpful. 
If a user has difficulty finding information he will 
seek other software or computer systems. 


Often points of "confusion develop when a 
user loses track of who is doing what, when, and 
where. For instance, when describing interactive 
dialog it is important to indicate when the user is 
to enter dialog and when the computer outputs 
dialog, i.e. who's turn is it. The relationship 
between several software functions can be lost 
through all the detail in each, thus a summary of 
the functions showing or describing their relation- 
ship is helpful to a user. 

One technique to keep the perspective is to 
ease the user from a more global concept and scope 
in on the particular item to be discussed. Other 
techniques include diagrams, outlines, pictures, 
and hierachical structured paragraphs. Of 
course, complete, but simple examples provide 
good perspective with user related topics. Such 
complete examples tie together many subtask con- 
cepts so the user can see the big picture. 

The user's primary concern is to perform 
his task WITHOUT getting involved. 


A useful restaurant adage is "if it looks 
good it tastes good. " This also applies to a user's 



BOX 1579, PALO ALTO CA 94302 

document and I'll never have to eat my words. This 
approach is done by all the big computer companies. 
They publish great looking documents but all too often 
once you get into it the user's document is little 
better than useless. For example, one manufacturer 
published a very pretty 300 page user's reference 
manual without an alphabetical index but the system 
sold well (boy ! did the users ever complain). 

Appearance is effected by the page layout 
such as the spacing around and between paragraphs, 
the type font such as script for "smooth" appearance 
or bold for mandatory and easily recognizable items, 
and page titles. Bold face font may be used to 
emphasize key words. Round cornered heavy plastic 
tabs are impressive though expensive. Even more 
expensive is the art work done to illustrate and 
diagram user information but it can appear most 
effective. One of the most effective ways to create 
a useful appearance is to use color; although it is 
expensive. Colors have psychological effects on 
users. For instance, the error description section 
of a user manual might have a red tab. On the 
other hand a tab for the index may be yellow while 
the sections describing the user commands to 
execute useful functions might be green. Along this 
line, the printed page stock could be colored. The 
appearance of colors dictates how the document 
is to be used. 

If the format appearance is done well a tech- 
nical document, such as a computer user's guide, 
can increase comprehension. This is based on the 
fact that most technically oriented people tend to 
respond to ordered, sequenced, and logically 
arranged material. 

Technical Information Structure 

The following is a guideline for structuring 
written user information describing a subroutine, 
program, subsystem, or computer system. 

Paragraph 1. Tell why the user would want to use 
this item of the system/software. 

Paragraph 2. General form or format of this item. 

Paragraph 3. What are the results of using the item. 

Paragraph 4. Typical simple example with 

Paragraph 5. Simple description of the most 
straightforward use of this item. 

Paragraph 6. More detailed description of how to 
use this item. 

Paragraph 7. Discussion of problems, limitations, 
and other pitfalls the user should 
guard against. Discussion of what 
to do if these problems should occur. 

Paragraph 8. Description of special and more 
complex uses of this item. 


You may well realize by now that these 
documentation ideas for users also apply to many 
technical documents, including program documen- 
tation. By keeping these eight elementary concepts 

in your mind you too can be a success with good 
user documentation easily and simply. 

If we are really to meet a user's needs with 
our computer systems shouldn't we write the user's 
documentation before we do the design and develop- 
ment of our software/systems ? 

B. the User: 

—is looking for solutions 
■needs a reference 
-normally uses this document 
has a complex task 
-finds a user problem 
- has a special task 



Then he will seek one of the 
following levels 

Why use 

General form of the solution 

Description of results 

Typical example 

Simple description 


Problem and alternatives 

Special uses 

Different Needs Dictate Different Uses 


Mecham, Douglas J. 
You and Written Information 
Technical presentation, HP3000 Users Group 
Meeting, May 1974, Chicago, Illinois. 

Melby, Michael P. 
Written Communications: A System Approach 
Human Factors Soceity Bulletin, Page 1-3, 
December 1977, Volume 20, Number 12. 

Pacific Printers PILOT 
Monthly magazine, Richard Zimmerman, Editor, 
M. L. Droubley, Publishers, 583 Monterey Pass 
Road, Monterey Park, California 91754. 

How to Use Graphics and Tables, Booklet 
From Interpreting Graphs and Tables by 
Peter H. Selby, Copyright 1976 by John Wiley 
and Sons, Inc. 

Himstreet, William C. 
Getting Your Words Worth 
Talk, Association for Systems Management 
Los Angeles Chapter Meeting, 1972. 



BOX 1579. PALO ALTO CA 94302 

Birkwood, Ilene 
The Technical Writers Survival Kit 
Hewlett-Packard, Technical Paper given at 
HP3000 Users Group Meeting, Seattle, Washing- 
ton, September 1977. (Excellent reference.) 

O'Hayre, John 
Gobbledygook Has Gotta Go 
Bureau of Land Management, U.S. Department 
of the Interior, U.S. Government Publication, 
1966, 0-206-141. (A must to read. ) 

Journal of the HP3000 Users Group 
(HP3000 Users Group Newsletter) 
Hewlett-Packard, Santa Clara, California. 

65 Notes 
Richard J. Nelson, Editor 
2541 W. Camden Place, Santa Ana, California. 


I wish to acknowledge three people who 
have assisted my preparation of this topic. First is 
my secretary, Lynda Schenet, who is a wizard 
with the typewriter plus spelling. Then there is 
Mr. Richard Nelson who inspires me to get- this 
kind of job done. The third person, Mr. Robert 
Barsalou, is from several years past who laid the 
foundation for these topics. 



BOX 1579, PALO ALTO CA 94302 


Richard J. Nelson 
Editor-Publisher PPC Journal 

2541 W. Camden Place 
Santa Ana, California 92704 


The fact that you read the title above and 
have gotten this far indicates that you would 
like to do some writing, have some ideas to ex- 
press, or are presently involved in writing, ed- 
iting, or publishing. The material that follows 
is an amateurs viewpoint, and the only justifi- 
cation I have for writing this paper is a suc- 
cessful newsletter(l). A newsletter is a fre- 
quent, less formal method of publishing timely 
information for a fairly narrow readership. 
Newsletters are periodicals that offer fast re- 
sponse compared to a magazine that is typeset 
and requires three months or longer leadtime 
from idea to print to reader. The Editor is re- 
sponsible for the content of the newsletter and 
the Publisher sees that the Editor's material is 
reproduced and delivered. For amateur newslet- 
ters the Editor and Publisher are often the same 
person. This paper will provide an overview of 
the task of editing and publishing a club news- 
letter. Specific tried and successful techni- 
ques will be included and some references provi- 
ded. This paper is not intented to cover the 
fields of graphic arts, technical writing, or 
printing, which would require several books. You 
should, however, be able to understand the prob- 
lems and have a base from which to start a news- 
letter if you are so inclined. 

Should I Edit a Newsletter? 

The motivation to become famous by circu- 
lating reams of world shaking information is 
within us all. In the real world, however, Edi- 
tors seldom become famous, or even well liked. 
A few do, but don't be misled into taking on an 
editing task if you don't like hard work and 
little recognition. You should plan on at least 
a year of editing if you are going to do it at 
all. It takes that long to get into a produc- 
tion mode and build a readership. Editors are 
not necessarily writers; they often write the 
material they produce, but they do not create 
the material, except for the first few issues in 
order to get started. There are exceptions * and 
there are many successful newsletters (success- 
ful here means profitable) which reflect the op- 
inion of an Editor knowledgable in a specific 
field. In the computer hobby field you will be 

spending your time stimulating, organizing, pre- 
paring, and printing. Computers are technical 
and you should have some technical or other ex- 
perience, such as software or applications, to 
draw upon for deciding the content of the news- 

In deciding if you should edit a newsletter 
you should be notivated by a desire to serve a 
cause, such as a club, or to make a contribution 
to the field by providing needed information not 
readily available to the intended reader. If 
fame and fortune is your goal, don't take on a 
club newsletter except as a training experience 
for something more profitable later on. Be pre- 
pared to work mostly alone and independently. 
The mean time to burnout for your helpers will 
be measured in weeks and seldom lasts longer than 
two issues. If you haven't been too discouraged 
so far, let's get into some specifics. 

What Do I Publish? 

A club newsletter will usually have some 
specific goals, such as announcing meetings and 
reporting on activities. Unless you are taking 
over an existing publication that is well estab- 
lished, you will generally have a great deal of 
freedom on what type of material you publish. 
Success, as defined in this paper, is having 
more good material for each issue than space per- 
mits, and having a continuous growth and demand 
for what you produce. A club newsletter serving 
a hobby readership must support itself in most 
cases. You must provide material that people 
will read. In the United States we have struc- 
tured our communications media in such a manner 
that we try to make it easy for the reader. News- 
letters can't waste space with wide margins, 
large type, and profuse illustrations. News- 
letters usually have a small readership and lit- 
tle advertising. If the material in the news- 
letter is not needed (technical) or interesting 
(well written and illustrated), it will not be 
read. Deciding on what to publish is what makes 
an Editor successful. 

The first task is to identify the purpose of 
the newsletter. If the club is dedicated to 
computerized chess, the newsletter style and 



BOX 1579. PALO ALTO CA 94302 

content will be different than if the club is 

dedicated to small business systems. The chess 

enthusiast is interested in a very specific, very 

technical subject involving great detail. The 

business system user will be interested in the 

system and what other people are doing with it. 

Most club newsletters are of the latter type, 

and one of the keys to success is to recognize 

the people aspect of a highly technical activity. 40 publications a month for useful material for 

not from a specifically acknowledged source has 
probably been prepared by the Editor, or staff 
writer (not too common with club newsletters). 
Few authors will wish to remain anonymous. 

Events, products, news, etc., requires that 
the Editor be well read in the many publications 
related to the newsletter topic. Scanni-ng 30 to 

For a computer club newsletter it is inappropri- 
ate to get too involved with non-technical top- 
ics outside of club social activities. Topics 
involving religion, social reform, etc., should 
be included only if the computer plays a role, 
and the people involved are recognized in their 
field or are part of the club membership. The 
decision of what to publish should involve a sim- 
ple formula or mix. Various categories can be 
made; they will usually include: 

A. club activities 

B. human interest/applications articles 

C. feature technical articles 

D. reporting of events, products, news, etc. 

Club activities will include meeting ann- 
ouncements, programs, officer changes, or any 
item concerning the club that the reader should 
know about. This information, like all infor- 
mation in the newsletter, should be easily found 
by using a format common to all issues. Format 
and organization will be discussed later. 

Human interest and applications articles 
should involve club members if at all possible. 
Computers are interesting, but people using com- 
puters_are more interesting. A newsletter T s . 
usually of the general interest type, as mention- 
ed above with the business system example. A 
construction or assembly article is much more 
interesting if Mary Ann describes her own ex- 
periences along with the useful technical de- 
tails the article provides. This encourages the 
reader to get involved, because he or she relates 
to Mary Ann. How other people are using their 
machines, or the justifications they made for 
buying their machines, are topics that readers 
will want to know about. These types of articles 
will require special effort on the part of the 
Editor to encourage, nurture, and coax into 

Feature technical articles will form the 
real backbone of the newsletter. You will have 
to know who is doing what in the club to draw 
upon the members technical expertise. As Editor 
you may have to write the article after getting 
the information from the actual source. The per- 
son providing information gets the by-line, not 
you. The Editor rarely signs his name to arti- 
cles even if he creates and writes the material. 
The reader can assume that all material that is 

this category is not uncommon for an Editor. 
These publications may be freebees, or traded 
for being on your mailing list. If your news- 
letter is successful and contributes ideas, other 
Editors will want to exchange publications. 

What Format Shall I Use? 

The format of a newsletter, as the term is 
used here, is the layout of the material on the 
page. Margins, number of columns, typesetting, 
headings, and use of character size, upper and 
lower case, etc. A newsletter could be printed 
on a microdot, or it may be a simple page-wide 
column typewritten piece. Neither of these two 
extremes are recommended. Avoid special type 
styles, and DO NOT USE ALL UPPER CASE for text. 
It is unfortunate that most computer printers 
are not designed for human beings to read, but 
this is changing as computer and peripheral manu 
facturers realize that computers process text and 
the output should be readable by humans. If pos- 
sible, type the text and use computer printed 
material for program listings and tables. The 
tradeoff is, of course, readability versus ac- 

The choice of a format will depend on what 
equipment is available to you ■, how much informa- 
tion you want to squeeze onto a page and, to some 
extent, the readership you hope will read the 
newsletter. Gather at least a dozen different 
newsletters and study them. Try to get as many 
different kinds as you can. Look over the for- 
mats and observe the following: 

a. how wide are the margins? 

b. how many columns? 

c. is typesetting used? 

d. what type styles are used? 

e. is colored paper used? 

f. what kind of illustrations are used? 

g. how many, what percentage of space, for pho- 

h. are the pages bound (folded) or single 

i. is it three-hole punched? 
j. what is the content, and readership of the 

As you study various newsletters and answer 
the above questions you will be able to get a 
'feel' for the type of format that you will want 
to use. Here are my recommendations: 



BOX 1579, PALO ALTO CA 94302 

Use a two-column format as a compromise in 
paste-up convenience and readability. Do not 
right-justify; most studies show that right-jus- 
tification slows down the reader and, unless you 
are trying to make an 'image' of being a so- 

3l«S!*H h I 9h 3 Ua I 1ty p1ece ' the extra Production 
effort detracts from subject content. Reduce 
standard typewritten (10 or 12 characters per 
inch) produced text to achieve at least 1,200 to 
1,500 words per page. Start page numbering from 
Page 1 which is also the front cover. Leave a 
wider left margin on odd-pages and right margin 
on even-pages for binding or hole punching. Make 
top and bottom margins equal. Place page num- 
ber, newsletter name or logo, volume and issue or 
date, on each and every page. This should be 
standard practice for everything printed today 
because of readily available photocopying. Most 
people want to know the source of useful infor- 
mation and each page copied should have that in- 

Masthead . At the top of Page 1 of each 
newsletter you will see the name and other infor- 
mation relavent to the newsletter. The masthead 
may be a simple name or it may be an elaborate 
design. The following information may be part 
of the masthead: 

1. publication name 

2. short description or explanation of the sub- 
's , J SJ» scope » or P ur Pose of the newsletter 

<5- LUGO 

4. issue identification by date, volume, issue 

b. cost 

6. copyright symbol 

. Time. The readers time may not be an ob- 
vious consideration in choosing a format A 
technical publication is usually read twice; 
initially, and later as a reference to obtain a 
specific part of the information. The layout and 
format, and especially the titles and illustra- 
tions, should be chosen carefully for descrip- 
tive accuracy and visual association. The effec- 
tiveness of the newsletter as a reference is di- 
rectly in the hands of the Editor. Time is also 
saved by utilizing the reader's familiarity with 
the newsletter. A newsletter that is consistent 

ITU ]! SU V° i ? S i e 1s easier t0 P roduc e and 
read if a formal format is established. 

per inch, column width, reductions, etc. A tyD- 
ical layout is shown in Appendix A. The figure 
shows the paste-up page and its dimensions alone 
with the print columns and margins. The formal" 
format should include the specifications for the 
newsletter and the paste-up. Reductions may be 
chosen as those used on the Xerox 7000. Reduc- 
tions #2 or 3 are good. #3 is especially prac- 
tical because the back of unusued, 132-column 
computer printout sheets (ll"xl4") make good 
paste-up sheets. The #5 reduction (61.5% of ori- 
ginal) is the limit that any reader will toler- 
ate and should be well justified before using 
Most will find the print too small for casual 

Preparing the Text . After all of the above 
preparations have been made, the actual news- 
letter can be started. The many aspects of wri- 
ting cannot be covered here except to provide a 
not-so-obvious truism; you learn to write by 
writing. The mechanics of getting from idea to 
typed text varies considerably from the Editor 
who only has pen and paper to the Editor who 
writes, edits, and types his copy. Unless you 
are an excellent speller, have someone else read 
your copy before starting the cut and paste 

Typing. The ideal typewriter is the IBM 
Correcting Selectric. The advantages of being 
able to use different type elements (balls) 
gives you a little more freedom in being able to 
express complex technical ideas in written form 
The newsletter material is single-spaced and 
typed the column width as defined by the format 
specifications. This means that a single column 
typed on 8V'xll" sheet will fit with space in 
the margins for notes. Column length can be 
thought of as continuous from beginning to end 
If you make a major error and wish to retype a 
single line or more, just make a notation in the 
margin to cut out during the paste-up stage. 

Formal Format . Write down a sketch of the 
layout that will define your format. Keep it 
handy so you may use it when you assemble the 
newsletter. The mechanics of the newsletter 
should be clearly defined and executed as a mat- 
ter of habit so you can be concentrating on the 
details of the content. As Editor/Publisher you 
may often have to be proofreader, illustrator, 
artist, etc. Once you prepare your formal lay- 
out you won't have to think about margins, lines 

Masthead. Your masthead will be the same 
for each issue except for Page 1 text and date, 
volume, number, etc. I suggest that you prepare 
a master page with all the material that won't 
change from issue to issue, and print 100 copies 

fn^D? 396 / 11 ^ 6 ™ 6 as a s P ecial Paste-up sheet 
for Page . This method insures a consistent 
print quality for your image-masthead. The idea 
approach is to prepare a paste-up sized master 
and print 100 copies of it for Page 1. The lar- 
ger sized plates and press required to do this 
makes the cost beyond the normal club resources. 

. Photographs. Photographs cannot be used as 
originals for making plates for a printer to 
print. Printing involves putting ink on paper. 
There is only white paper and black ink. There 
is no gray or shades of black! A photograph has 



BOX 1579. PALO ALTO CA 94302 

many shades of gray and must be screened to pro- 
vide a photograph (positive) that is made up of 
black dots that are either touching (black) or 
further apart (appears as gray) to give the ap- 
oearance of shades of white to black. The size 
and spacing of the dots control the detail of 
the printed image. If the dots are small and 
close together, a higher resolution (number of 
lines distinguishable per inch) is possible. For 
low cost printing, especially paper plates, pho- 
tographs should be screened 65 or 85 lines. If 
you take the photographs, or print them yourself, 
you should make them a little on the light side. 
Mount them on a piece of cardboard and take them 
to a graphics arts studio that will screen the 
photos with a 65 or 85 line screen. Ask for a 
screened positive. The originals will be re- 
turned to you so you can return them to the au- 
thor. Photographs are easy to include in the 
newsletter, and once you have found a quality 
source for the screening, you should have some 
photos in every issue. 

If your screened photograph is a Kodak PMT 
and you save your paste-ups for reprints at a 
later time, I suggest you wash them if your gra- 
phics arts studio does not. If they are not 
vashed, they may turn yellow in a few months. 
Simply place the prints in the full bathroom 
sink and slowly run cold water to rinse them 
for 10-15 minutes. Pull out, place on the 
Dathroom mirror and use a clean window squeegee 
to remove excess water from both sides, and 
then hang to dry. 

The Paste-Up . If the newsletter page is to 
be reduced, the camera (or. Xerox) copy must be 
larger. For example, this paper was prepared on 
a paste-up sheet with page size 10.24" x 13.25" 
which, when reduced to 83% of paste-up is SV'xll" 
Don't get confused with the ratios; professionals 
rtork in picas, etc. Having an engineering back- 
ground I simply use a ruler maked off in tenths 
Df an inch and use a calculator. Two factors 
should be noted on your formal format sheets; one 
to determine paste-up size from newsletter size 
(greater than one), and one to determine news- 
letter size from paste-up size (less than one). 
The table below illustrates the relationships. 
The two factors are reciprocals of each other. 
The Xerox 7000 reduction #3 is used: 

It is not practical to type directly on the 
paste-up sheet even if you have a wide carriage 
typewriter. The common practice is to cut and 
paste the material to the paste-up sheet. Rubbe 
cement is often used but I do not recommend that 
you use it. The best method is to use a hand 
waxer, or Scotch #810 tape. 

Paste-Up Techniques . The equipment used to 
perform the newsletter paste-up is described in 
the following section. You have typed your arti- 
cles in the column widths as determined by your 
format, and are ready to start your paste-up. 
Commercial paste-up sheets are available from 
printers or Graphic Arts suppliers. They will 
have light blue lines that serve as guides to 
attach the prepared typed text. You may not be 
able to find ones to suit your needs and will 
have to improvise. If a 77% reduction is used, 
11x14" computer sheets work fine. Cuff off the 
holes at the end and prepare a stack for a years 
use. Add your column guidelines by using a light 
blue pencil and a cardboard template that has 
cutouts and a block of wood for a handle. The 
side margins will not be equal, so mark the top 
of the template ODD PAGE with the wide margin at 
the left; turn the template 180° and mark the 
top EVEN PAGE with the wide margin to the right. 
You may mark all pages the same way and rotate 
them as you use them, or you may mark each page 
or E in sequence. The latter method takes 
longer, but helps prevent errors if a non-reduced 
page is used. Use one system and be consistent. 
The idea is to establish a system for the first 
issue and have the mechanics follow the system 
so you can concentrate on composing the page. 

The paste-up sheet will be reduced by Xerox/ 
or camera by the printer. When a column of ty- 
ped text is cut by papercutter I recommend leav- 
ing about 1/16" white space on each side of the 
column. This allows the printer to be able to 
remove any shadow if he is using an electro- 
static plate maker. If you reduce your paste-up 
on a Xerox machine, I suggest you place a soft, 
thick, white ink blotter, larger than your paste- 
up sheet, on top of it. Press on the Xerox rub- 
ber cover to keep the paste-up sheet as close to 
the glass as possible - the ink blotter paper 
(or dozen sheets of newspaper) helps distribute 
the pressure to keep the whole sheet flat. This 
will reduce the shadows that may be picked up by 
the non-parallel light source in the machine. 

Paste-Up Reduction 
Sheet % of 

for News- 
letter Size 




Fable 1. Paste-Up vs Newsletter Page Sizes 

A useful technique for cutting the typed 

M ... ,. text with a papercutter is to place a large white 

Multiplier sheet Qf paper Qn the table under the blade . This 

for paste-up reflects overhead light, and your hand placed 
size - above the text being cut provides a light con- 
trast that shows a shadow of the bottom edge of 
— the papercutter. A precise cut can be made usinc 
this technique. A single line of type can be 




BOX 1579, PALO ALTO CA 94302 

removed without difficulty. An Xacto knife is 
useful, but the papercutter method is faster and 

Attaching the cut text or photos to the 
paste-up sheet may also be done using a small 
sliver of Scotch Magic Tape #810. Using a tape 
dispenser, a pair of scissors, and tweezers, cut 
the tape as follows: pull tape out 2", cut 1/16" 
off the end to remove the jagged edge made by 
the previous cut of the dispenser. Hold the 
'floating' end of the tape with the tweezers. 
Cut a 1/8" length with the scissors. Hold with 
tweezers and tape the photo, etc., to the paste- 
up sheet. This small amount of tape won't be a 
problem for the plate maker (larger pieces will 
reflect too much light), and using the tweezers 
to slip under the 'pasted' material it is easy 
to move if required. 

Using a hand waxer is the best way to assem- 
ble a newsletter. The paste-up sheet may be 
waxed, but it is best to wax the material being 
laid down on the paste-up sheet. A stack of 
newspapers, cut in half to reduce table space, 
makes a good work area to roll the waxer over 
the text, etc. (upside-down). The newspaper 
is discarded once there is wax on it to avoid 
the disaster of getting wax on the text side of 
the paper. 

Equipment . Preparing a newsletter requires 
some basic equipment which is essential if you 
expect to survive even one year. The first piece 
of equipment you should have is a large, good 
quality papercutter. Spend the $20 to $35 to 
get one that has at least a 12" blade. Next is 
that marvelous gadget - the waxer (2). A waxer 
is a small roller with a heated reservoir of hot 
wax which rolls a 2" wide layer of a special 
sticky wax over anything. The paper sticks be- 
cause the wax has a sticky characteristic that 
holds paper to paper when a small roller is used 
to press the sheets together. Tweezers (Clauss 
#AA), Xacto knife, and large 18" ruler with 1/10" 
scale are important tools for an Editor doing 
his own paste-ups. 

Printing . 

Acceptable, low cost printing can be found 
in most larger cities. Southern California, es- 
pecially the Los Angeles-Orange County area, 
seems to offer the lowest cost printing in the 
U.S. One hundred copies of an, orig'irial costs 
$2.12, tax included. iwo k hundred copies, $3.18. 
This offset printing process uses an electro- 
static process to make a paper plate. Check with 
local printers to see what is available. Visit 
tfieir shops and see what equipment they have. If 
you reduce your paste-ups free on a friendly em- 
ployers Xerox 7000, you will have a continuous 
problem of varying quality. Usually the printer 

has a camera capable of handling most paste-up 
sizes and he will make the plate from the paste- 
up. I want a permanent file master and pay $2.0C 
per page for a reduced PMT positive print for 
each page. The printer makes his plates from 
the PMT, This has the additional advantage of 
being able to add full size titles waxed to the 
PMT to give larger headings, etc. It does add 
an extra day to the production process. 

There are many methods of producing the man> 
copies required to mail or distribute to your 
readers. Offset printing is best, Xeroxing or 
other photocopying methods, such as the 3M or 
Savin copier, may be used to obtain two-sided 
copies. Offset printing is lower in cost and 
provides better quality than photocopying. Ask 
the photocopy shop if they have a price and ser- 
vices list. Examine it - most likely it is 
printed! I once experimented with an issue re- 
produced on a Xerox 9200; the newsletter was 32 
pages and I had 2,000 copies run. The pricing 
was a little higher than printing; delivery time 
about the same for collated sets. Quality, how- 
ever, was inconsistent and I never went back. 
My friendly printer gives me his odd sheets (the 
'extras' a printer prints to insure having e- 
nough printed properly on both sides) which make 
useful material for correspondence, paste-ups, 

Collating . 

Most printers will collate your newsletter 
for an extra charge. If the newsletter is a 
club effort, you can do the collating at the 
same time you prepare the newsletter for mail- 
ing. A bookshelf makes a good collating device, 
or a homemade collator can be made. Cut slots 
for 9%"xll 3 s" sheet metal shelves, 1" apart, for 
a small 'bookshelf 9" wide. The height of the 
collator can be about 18" for a 16-shelf capa- 
city. The horizontal sheet metal shelves can 
be slid into their slots 2" for the bottom, 3" 
for the next, etc., going up. Five to sev,en« * 
sheets can be collated with about IPO sheets per 
shelf capacity. A wet sponge moistens the index 
finger to pull the sheets out, and a quick col- 
lation of 300 sets of a 14-page newsletter can 
be done jn«an hour. 

Printing the first page of the newsletter a 
different color gives it character and allows 
many issues to be stacked for storage and easy 
separation later. 

Mailing . 

One of the biggest problems is the mailing 
list and production of labels. I suggest the 
following scheme to maintain the mailing list 
and produce the proper labels for each mailing. 



BOX 1579, PALO ALTO CA 94302 

i Assign each member/subscriber a sequential num- 
Iber. Type his name, number, and address in three 
or four lines centered on an Avery label sheet, 
#5351, intended for Xerox copies. Write the 
member's expiration date in pencil on the lower 
right part of the label. The 5351 label sheet 
has 33 l"x2-3/4" labels stuck to a waxed sheet. 
Reserve the upper-right label for a large number 
to identify the sheet. When mailing time arrives 
scan each sheet to find expired members. Peel 
off their label and place on the back of the 
sheet. Take your box of master labels to a Xerox 
machine and have them duplicated onto another 
set of 5351 labels. Peel and stick the dupli- 
cate set to envelopes or self-mailer newsletters 
at a rate of 450-500 per hour. When a member 
renews, place his label back on the front. 
Changes are easily made. Do not make copies in 
advance; things change too fast. A mailing list 
of 3,000 has proven no problem -in maintaining in 
this manner. 

Another scheme that has worked with older, 
slower photocopiers, is to photocopy Page 1 which 
is also a self-mailer and contains the address 
information. The mailing list is typed on a 
length of adding machine tape with convenient 
spacing. Two slits are made in the original 
placed on the photocopier window. After each 
copy the tape is pulled through the two slits to 
show the next address. The sequential numbering 
of the names allows a quick check that the list 
was complete. Both methods have been used with 
success - never does a member not get an issue 
mailed to him. The Post Office may lose it, but 
you have one addressed to him each issue. 

Avoid sticking stamps if possible. It is a 
chore because the managers of our postal system 
have no concept of providing stamps in convenient 
form. Rolls are reasonable for various forms of 
mechanization, and go fast manually. Sheet 
stamps can be stacked five to ten in a stack, 
stapled at one edge, and cut into strips on the 
papercutter, leaving the whole sheet still atta- 
ched (for counting and control) with the cut 
stopping V' from the end. Tear off one stack 
strip at a time and apply using a sponge. The 
problem is that usually the required amount of 
postage is not a single stamp in roll form. 
Foreign postage requires extra time, so allow 
your extra charges to cover this expense. Send 
First Class to insure delivery — remember, a 
newsletter is timely. Foreign is sent Air MaiN 
Other Article - and requires a non-sealed clasp 

Obtain a mailing permit if you qualify. A 
mailing machine is OK if your club doesn't have 
to buy and maintain it. Mailing permit informa- 
tion, and mailing costs for various classes of 
mail, is wery difficult to get from the Post Of- 

fice, but give it a good try. Go to a Main Post 
Office after calling to make arrangements to talk 
to someone who handles business accounts. The 
usual window teller cannot help you much. Per- 
mits are reasonable in cost and allow you to 
print a cancelled logo on the envelope. Most 
newsletters are marked FIRST CLASS, dated mater- 
ial, and mailed with a permit. 

Finances . 

The club usually finances the newsletter. 
Costs for a year's operation are very predic- 
table if a fixed issue size, including attach- 
ments (such as a member list, index, etc.), is 
determined. Allow 20 or 30 issues for Editors 
exchange and overprint enough for those special 
packages of back issues an Editor will want to 
swap with other Editors. You should not have to 
worry about finances or getting checks to pay 
the printer. If you do your planning properly 
before your first issue is printed, you will 
save considerable time and frustration. Printing 
costs, label, envelope, and assembly costs will 
be small compared to mailing costs. Allow for 
supplies such as tape, rub-on labels, liquid 
paper, black pens, and blue pencils. Typing can 
be professionally done without financial diffi- 
culty if your mailing list is several hundred. 
Do not take on a newsletter that can't support 
itself. Advertisements will help, but for a 
small operation it won't contribute much and are 
essentially donations by local businesses. 

Conclusion . 

Editing and publishing a club newsletter 
caji be a rewarding and educational experience. 
Most people have no concept how time consuming 
it can be. The Editor will spend six to ten 
hours per page per issue if he does not do the 
typing, and his print density is 1,200 plus wordSi 
per page with illustrations and photographs. The 
pressure of working to a deadline can be too 
stressing for some people. It seems that you re- 
lax a few days after crashing out the last issue, 
then you realize that you are getting late for 
the next one. Your reward for taking on the 
task must be self- fulfillment of getting good 
technical information to your readers and the 
pride of doing each issue a bit better than the 
previous one. 

There are no special requirements or train- 
ing that will make you into a good Editor. A 
certain feel and interest in information gather- 
ing and dissemination is required. Editing, 
writing, and programming are similar; you really 
don't know if your efforts will be successful 
and rewarding until you try. 

See bottom of Appendix page for footnotes. 



BOX 1579, PALO ALTO CA 94302 


Paste-up sheet layout 



1" (.77) 


Column is 53 characters 
across @ 12 char. /in. 
This is 8.8 words per 
column. (6 char/word) 
Column length is 12 in. 
Six lines/in is 72 lines. 
Words per column is 
72 x 8.8 = 634. Words "^ 
per page is 1,268. 

Note: Typeset text has 
variable spacing and an 
average must be used if 
I a comparison is made. 





-0.9" (.69) 


.5" -j 







1" (.77) 


Odd page and front page, even page rotate 180°. 

Note: Numbers in parenthesis ( ) are newsletter page dimensions. 

This layout uses the reduction shown in table 1 of the text. It is a good 
place to start if you are doing your first newsletter. It is effecient in 
that 1200 words per page is still in a type size that is readable by most 
people. Make notes to formalize your layout to insuure a consistant newsletter. 

Footnotes . 

(1 ) The newsletter is the PPC Journal, formerly 
called 65 NOTES, a monthly publication of 
PPC, formerly called the HP-65 USERS CLUB. 
PPC is a dedidated Personal Programmers 
Club for Hewlett-Packard Personal Program- 
mable Calculators. PPC Journal readership 

is nearly 2,000. PPC members contribute 
$15 per year for the programs, programming 
techniques, applications, and hardware modi- 
fications information published in the PPC 
Journal . 

(2) Write Paste-Up Supply, 1113 Walnut Street, 
San Gabriel, California 91776, (213) 283- 
4610, for information on waxers and supplies 



BOX 1579, PALO ALTO CA 94302 


The True Computer ist 


Tom Pitt man 

P.O.Box 23189 

San Jose, Ca. 95153 

Several years ago, when the idea of a personal computer was still 
only a gleam in my eye, I made an observation about my work as a 
computer programmer. I suppose the same thoughts have occurred 
in the minds of painters, sculptors, composers and other artists 
down through the ages. I perceived that I was giving existence to 
something which had not existed before — I was creating ex nthtto, 
out of nothing. To be sure, most of the programs I wrote were mere 
copies or adaptations of other programs, and nothing new in 
themselves. And, the taxman to the contrary, there was no tangible 
substance to the work of my hands. But every once in a while I could 
stand back and look on my work and say, "See what I made!" 

I do not wish to quibble at this point with those who claim that 
nothing is truly a creation. As I said, most of the programs I write 
are merely copies or adaptations of some other programs. I am not 
talking about those. Nor do I particularly wish to quarrel with 
oehaviorists or social biologists who reduce every activity of Man to 
the effects of his environment or his genes. What I am getting at 
here is the particular feeling that only comes with knowing you have 
created something new. It is not quite the same as the feeling an 
expert technician gets in his craft, the sense of doing a job well. I 
have felt that often, and I continue to take a certain pride (if you will 
pardon my immodesty) in the high technical quality of my work. 

There is a difference between the technician and the artist. The 
technician is building to an existing pattern or plan; the artist is 
making a new plan. The technician has a standard by which to 
measure tiis work; the artist ts his own standard. Of the- 
technician's work you can say "He did (or he did not) meet the 
requirements of the specifications"; of the artist's work you can only 
say "Ahhh!" or "Yecch!" 

I wish the boundary between the technician and the artist were 
that clear-cut. It is seldom so. What the artist creates is, whether 
he likes it or not, subject to technical criteria. If he is painting a 
portrait or a landscape, you can apply the purely technical 
judgements to it of whether or not it adequately conveys an image of 
the subject. You can even determine if the paints have been mixed 
correctly, or if they are likely to deteriorate and change color with 
age. Is the perspective and lighting believable? Into this technical 
fabric, however, is woven the art, the quality that makes Dttrer great 
and Cranach so-so. 

I am a programmer, not a painter. It is much harder to see the 
"art" in a computer program. Donald Knuth sees it and the title of 
his monumental work on programming is "The Art of Computer 
Programming" [1]. I only hope nobody asks me to point to some 
program and say "this is art, not technique." Perhaps I am a coward 
and lack the fortitude to defend such an assertion. Perhaps there is 
no defense and I dislike making indefensible assertions. No matter. 
.Vly point is that the assertion can be made, and that it has meaning 
(at least for most of us). 

I raised the issue in connection with a certain feeling I got as I 
reviewed my work, when I saw it as a creation. You see, in that 

instant I as a Christian thought I could feel something of the 
satisfaction that God must have felt when He created the world: 
"And God saw every thing that he had made, and, behold, it was very 
good." [2] If man is, as Christians believe, created "in the image of 
God" 1 3], then perhaps I had learned something about God. In this I 
have a definite advantage over the painter and the composer: I can 
create something that will interact with me, as man interacts with 
God. So far it is a strictly intellectual interaction, and for the most 
part very predictable — my creation does what I programmed it to 
do, which is (usually) what I had intended. I consider my insight to be 
only a pale reflection of the devine, but... 

Yet in my relationship to the computer and to the programs I 
write for it there is another dimension, wherein lies a very grave 
danger. The danger is that I will lose my sense of perspective, and 
forget the relationship between God, myself, and the computer. 

Theodore Nelson in his popular book, Computer Lib , refers to a 
"computer priesthood." [4] By this he means that the computer 
technology has built up around itself a kind of mystery religion or 
gnosticism, with the computer professionals acting as the priests of 
that religion. Gnostic religions in history have had a body of secret 
knowledge (the word "gnostic" is derived from the Greek word for 
"knowledge") which only the insiders have access to. just from a 
technical point of view this was, and continues to be, a very real 
problem. The use of computers in our time requires such a heavy 
t echnol o gicaf background that- outsiders are locked out. 

To be introduced into this computer gnosticism has, until only 
recently, required skill with a soldering iron (not just "Which end is 
hot?" but the proper way to apply solder to microelectronic circuits, 
special soldering tools, etc.), understanding of how to read resistor 
codes and the cryptic markings on integrated circuits, proper work 
habits for protecting delicate MOS components from static 
electricity, and the ability to decipher inadequate instructions and 
third-generation xerox drawings. With a small but increasing 
number of exceptions, the novitiate computerist is required to know 
(but is not given instruction in) binary, octal, and hexadecimal number 
systems, machine language programming, real-time I/O control, 
ASCII code translation, and memory management. Unless he is 
willing to remain in the outer circle playing games that someone else 
wrote, the new gnostic is compelled to learn a foreign language 
something like Latin (BASIC) and be able to construct in that language 
esoteric hymns called Loops, Subroutines, Conditionals, 
Assignments, Input/Output and Data statements. I can assure you 
that the language requirement is not going to disappear for many 
years. We may get new languages, but that will only mean that the 
insider must know more, not less. 

Until the advent of personal computers in 1975, Computerism was 
successfully restricted to the elect who went through the necessary 
training and were employed in the computer departments of those 
corporations, educational and government institutions rich enough to 
be able to afford them. It was a closed society. With the advent of 



BOX 1579, PALO ALTO CA 94302 

the truly personal computer the ranks have been opened up to admit 
something over 100,000 new converts, but the careful observer will 
notice that it is still a closed society. The outsider is not really given 
much reason or help to join. Almost all of the magazine articles are 
directed to intermediates, not beginners. There are a very few 
books aimed at the novice (but not the totally uninitiated!), but they 
tend to get lost in the vast majority of books for the more 
sophisticated* By contrast, the number of Christian books aimed at 
the novice and uninitiated far exceeds those that require a significant 
background in theology. Computerism is still a closed society, though 
much larger than it was five years ago. 

So far I have been talking only about the phenomena of gnostic 
computerism, the appearances of the society. There is a deeper 
level that is much more serious. It is where we, the practitioners of 
this arcane art, begin to believe as we act. When we actually come to 
depend on the Computer to solve all our problems and to resolve even 
the mysteries of life, we have taken the final step toward making the 
Computer our god. At this level, opening up the secrets of 
computerism to all comers makes no difference at all; whether the 
Computer is a gnostic god or an evangelistic god is of little 
consequence, because we are talking about individual attitudes 
towards the machine. That is you and me, not "us" or "them." 
Institutions are formed from the aggregate of individual attitudes 
and beliefs. 

Before I get into what the Computerist believes, let me say 
something about the nature of belief. All of us prefer honesty over 
deceit (at least in the other person), kindness over malice (when all 
other things are equal), and so on. In the words of Mammy Yokum, 
"Good is better than evil, because it's nicer." [5] But when it gets' 
down to cases, with two of you out on the raft in the middle of the 
Pacific and food for two days, the choice of who eats the food and 
who dies of starvation depends on what you really believe. Or closer 
to home, it's rush hour and you are late to work; there is a long line 
of cars behind yours and you come to an intersection where a little 
old lady wants to cross the street you are on. This is where belief 
affects our lives. Your faith — your religion, if you will — is what 
makes the decision when it could go either way, but for different 
reasons. Notice here that I am not using the term "religion" to mean 
the religious institutions of society. Many of these function only as 
social institutions with little or no effect on the lives of the 
adherents. Instead I use the definition that equates religion with 
whatever is foremost in a person's thoughts and actions. In this 
sense everyone has a religion: some of us are Christians. Others 
are hgoists, "Moneyists", Scientists, Marxists, Sexists or 
Computerists. The focus of our attention is our god. 

So what is it that the pious Computerist believes? Which way will 
his decisions go, when it gets down to the crunch? Let me list a few 
"articles of faith" that affect the every day life of the practicing 

1. The computer is more interesting than most people. I love to 
spend time with my computer. It is fun to write programs for it, to 
play games on it, and to build new parts for it. It is fascinating to try 
to figure out what part of the program it is in by the way the lights 
flicker or the radio buzzes. It beats dull conversation any day. 

2. It is most important to be sure the computer is properly taken 
care of. When it finishes its present task, I must drop everything 
and go start up its next task, or turn off the disk drives to save wear 
and tear on the moving parts. If there is a power failure, ftrst go 
shut down the peripherals (save the disk, turn off the paper tape 
punch, etc.), then come back and get a candle to light up the house. 

3. The computer will be a big benefit in all kinds of ways. It is 
not connected up yet, but "soon" it will control the sprinklers, serve 
as a fire/burglar alarm, maintain the Christmas card mailing list, 
control the stereo tape deck, monitor the central heating, maintain 
the inventory in the pantry, provide menu planning, remind us of 
important dates, write form letters to people we don't like, educate 
our children, and so on. The key concept is that all these things are 
in the very near future. The computer has not yet fulfilled these 
promises, but it will very shortly now. 

4. The computer needs just a little more (memory) (speed) (disk 

space) (peripherals) (fidelity in its cassette drive) (better BASIC) 
(newer CPU) (noise suppression on the bus) (debugging on this 
program) (powerful editor) (bigger power supply) before it can do this 
or that. 

5. The computer can make money for us on the side, and 
eventually it will pay for itself. 

6. Spending all this time on my personal computer will give me job 
skills that will improve my wage-earning ability and make me eligible 
for a promotion. 

7. The computer can be used in the company business to improve 

8. There is no need to buy this software package or tnat circuit 
board; I can design one better. 

9. Let's arrange our next vacation to coincide with the Computer 
f aire. Wow! what a way to spend a vacation! Then we can swing by 
these manufacturers and/or these stores and see the latest widgits. 

10. To stay on top of the field it is necessary to subscribe to all 
five of the "Good" computer magazines. The other six are merely 
repetitious or uninspired or are beholden to their advertisers; they 
are not worth the subscription price. But when a good one comes sit 
down immediately and do not get up until I've gone through the whole 

11. Never miss a club meeting. This is where it's at. The juicy 
little news bits, the how-to-fixits for the problem that has been 
bugging me for the last two weeks, hearing about the interesting 
things that I can do with my computer — that is the real thing! 
Besides, they might have some free software. 

12. Visit the local computer store at least once a week. They 
are always getting new hardware in, and sometimes you can pick up 
some good rumors or a new book. You never know when you might 
bump into an Interesting Person at the store. 

None of these claims, by themselves, are particularly wrong or 
indicative of misdirection. But taken as a whole, they reflect an 
attitude that the computer is, in Paul Tillich's words, "the ultimate 
concern." [6] The computer becomes the object of one's devotion, 
the provider of one's needs. The computer has become the Absolute! 
the god in one's life. It is the work of our hands and the image of our 
minds; are we going to let it become the Lord of our lives? 

I would like to take exception to some of the "articles of faith" 
listed above. I have been there and I know the attitudes behind them. 
But I also know what is wrong with some of them. 

1. Clearly the computer is a fascinating device. Its complexities 
are overwhelming. I said this at the beginning, and I do not deny it 
now. But being the product of our own imagination, the computer 
probably cannot exceed our own intelligence. It may be faster and 
more accurate. It may do some things (like play chess) better 
because of this greater speed and accuracy, but it can never give us 
true inter raction on a human level. I realize I am going out on a limb 
in saying this; roboticists will gladly point to claims that humans 
would never leave the surface of the earth. Those claims were 
obviously wrong; I may be wrong also. But I can point to the 
difference between a technician and an artist with which I began this 
essay: vVe may be able to build robot technicians, but not robot 
artists. In any case it is unlikely to happen in your or my lifetime. 
The computer is a toot, and we should recognize it as such. 

2. The computer is a delicate machine and as such it requires 
care and maintenance. It is relatively expensive (today) and abusing 
it is not economically reasonable. But it is still a machine; it is not as 
important as any human being. 

3. The computer is capable of many kinds of benefits. But 
honestly now, how many of those things do you think you will actually 
get your computer to do? How many people do you know or have you 
heard of whose computer actually does those things, or even just 
some of them? Have you considered the transducers and mechanical 
linkages required to give the computer a meaningful selection over 
your music collection? Do you have any notion of the software effort 
needed to implement a computerized calendar or heating control (I 
mean beyond what is more easily done without a computer)? Have 
you ever stopped to think how much manual effort is required to 



BOX 1579, PALO ALTO CA 94302 

maintain a computerized pantry inventory? Have you considered the 
massive data entry requirement to build a data base for a decent 
menu planner? If educators and computer professionals working 
with large government grants cannot make much headway in 
Computer Assisted Instruction, are you going to master it in your 
spare time? Don't get me wrong. Many of these things are practical 
goals for computer implementation, but most of us, working on it as a 
hobby, will not get very many of these exciting applications working 
in any meaningful way. 

4. In the microcosm, the need for a little more of this or that for 
the computer seems very reasonable. A few years ago the Sunday 
supplement of some newspaper reported a survey on the money 
wishes in America. The result was that the average American 
would be happy if he or she had $13.21 more. Parkinson's Law holds 
that expenditures will always meet or exceed income. The same law 
applies to computer memory, speed, peripherals, and so on. 

5. Several of the magazines are touting the economic rewards 
potential in the personal computer. They are wrong. Suppose you 
did get a little extra money on the side computing bowling handicaps. 
What is to stop the local bowling alley from seeing the profit potential 
and buying their own computer? Maybe you will sell a neat game to a 
national distributor, but how many others are pushing games to the 
same distributors? Anyway, have you invented any neat games? 
Obviously some computers pay for themselves (mine does), but far 
more only promise to do so. 

6. Right now there is a shortage of people with microprocessor 
experience. Five years ago there was a shortage of COBOL 
programmers, but there is a glut now. People with drive and 
dedication, who make themselves valuable to their company, have no 
trouble finding work. People who stay up late nights on their own 
projects and are too tired to give the boss an honest day's work, who 
spend hours on the telephone ordering parts on the company bill for 
their personal computers, will find trouble finding and holding any 
kind of job. 

7. Yes, computers have improved the profitability of some 
companies. More often they have only promised to do so. You do not 
replace a bookkeeper with a computer; you give her a raise and call 
her a computer operator. It is still some time before we will see 
much business software available and usable. 

8. This one is subtle. Of course you can design one better. The 
computer is above all things an optimist's machine. But you won't. 
You don't have the time to get around to it, or it seems to have some 
problems when you do get it built: it never seems to work exactly the 
way you planned. 

9-12. by now the computer has moved out of the den and into the 
rest of your life. It will consume all of your spare time, and even 
your vacation, if you let it. It will empty your wallet and tie up your 
thoughts. It will drive away your family. Your friends will start to 
think of you as a bore. And what for? 

A few years ago I was doing some technical writing for a major 
electronics firm, and I had described some control circuit in terms 
such as, "the device provides such-and-such functions..." One of 
the people assigned to review my work reprimanded me: "Only God 
'provides 1 , circuits just..." I can no longer recall the exact details of 
the exchange, but I have not forgotten the message. God provides. 
Electronic circuits in general, and computers in particular, are not 
God; they provide nothing. They may be works of art, a beauty to 
behold, but in the final analysis, they are tools and they perform 
certain functions at our command. If we forget that computers are 
only tools, perhaps we will also forget that people are not tools, 
when we know Who is God, we also know who we are, and what 
computers are. In the words of the Second Commandment, 

Thou shalt not make unto thee any graven image, or any 
likeness of any thing that is in heaven above, or that is in the 
earth beneath, or that is in the water under the earth: Thou 
shak not bow down thysetf to them, nor serve them: for I the 
LORD thy God am a jealous God. [7] 





Donald E. Knuth, The art of computer P rogramming. Reading, 

iviass: Addison- Wesley 1973. 

Genesis I 31. 

Genesis I 27. 

Theodore H. Nelson, Computer Lib, p.2. Chicago: 1974. 

Al Capp, Lit Abner Sunday comics approx. 1974. 

Paul Tillicn T&ynamtcs of Faith, New York: Harper 1957. Most of 

this book Is nonsense, but Tillich does have a good 

understanding of what constitutes idolatry. 

Exodus XX 4,5. 



BOX 1579, PALO ALTO CA 94302 


James S. Albus 

(c New World Books 
4515 Saul Road 
Kensington, Maryland 20795 


Where are computers taking society? 

Will industrial and business robots 
lead to: 

Orwellian dictatorship? 

Jef f ersonian democracy? 

A new aristocracy based on robot 

Who will own these machines, and who 
will control the economic wealth and 
political power they will create? 

The great challenge of the coming 
industrial revolution will be to develop 
an economic system wherein prosperity can 
be achieved without waste, affluence can 
be made compatible with the limits to 
growth, and personal freedom can be pre- 
served and enhanced in a world where 
most economic wealth is created by auto- 
matic machines. 

Peoples ' Capitalism is a plan for 
meeting this challenge. It is a formula 
for a new economic order which might 
best be described as Jeffersonian democ- 
racy for the post-industrial era. 

The details of this plan will be 
described and a program for political 
action will be presented whereby Peoples' 
Capitalism could be introduced into any 
country in the world by the year 2007. 

Epilogue to Scarcity 

These are revolutionary times. 
Changes as profound as those resulting 
from the invention of agriculture or the 
domestication of wild animals are rushing 
us toward a new world. The human race 
is now poised on the brink of a new in- 
dustrial revolution which will at least 
equal, if not far exceed, the first in- 
dustrial revolution in its impact on 

mankind. The first industrial revolu- 
tion was based on the substitution of 
mechanical energy for muscle power. 
The next industrial revolution will be 
based on the substitution of electronic 
computers for the human brain in the 
control of machines and industrial 

From the beginning of human exis- 
tence, mankind has lived under the 
ancient biblical curse: "By the sweat 
of thy face shalt thou eat bread, till 
thou return unto the ground." Before 
the invention of the steam engine, 
virtually all economic wealth was 
created by the physical labor of human 
beings, assisted only by their domestic 

The first industrial revolution 
only partially lifted the ancient 
curse. Yet, even this partial reprieve 
had profound consequences. In all the 
thousands of centuries prior to the 
first industrial revolution, the human 
race existed near the threshold of 
survival, and every major civilization 
was based on some form of slavery or 
serfdom. Yet a mere two centuries 
after the introduction of steam power 
into the manufacturing process, slavery 
has become little more than a distant 
memory for the citizens of every major 
country. Today, a large percentage of 
the population of the world lives in 
a manner which far surpasses the wild- 
est Utopian fantasies of former 

There is good reason to believe 
that the next industrial revolution 
will change the history of the world 
every bit as profoundly as the first . 
The application of computers to the 
control of industrial processes will 
bring into being a new generation of 
machines ; machines which can not only 
create wealth unassisted by human 
beings, but which can even reproduce 
themselves at continuously decreasing 



BOX T579, PALO ALTO CA 94302 

costs. The potential long-run effects 
of this event are twofold: First, it 
will allow man to free himself from the 
dehumanizing demands of mechanization. 
The self -regulating capacity of computer- 
controlled industries will render it un- 
necessary for people to structure their 
lives around daily employment in factories 
and offices. The first industrial revol- 
tion drew people away from the land and 
concentrated them in urban industrial 
communities. The robot revolution will 
free human beings from the pressures and 
congestion of urbanization and allow 
them to choose their own lifestyles from 
a much wider variety of possibilities. 

Second, the introduction of the 
computer into manufacturing has the 
potential for removing material scarcity 
from the agenda of critical human prob- 
lems. The technical feasibility of fac- 
tories and industries which can operate 
unattended and reproduce their own essen- 
tial components implies that manufactured 
goods may eventually become as inexpen- 
sive and unlimited by process complexity 
as the products of biochemical mechanisms 
in living organisms. Increased efficiency 
and flexibility of substitution between 
materials and processes could render 
currently projected shortages of fuel 
and materials largely irrelevent to the 
21st century. 

Unfortunately, the present economic 
system is not structured to deal with the 
implications of a robot revolution. 
There presently exists no means by which 
average people can benefit from the un- 
precedented potentials of the next 
generation of industrial technology. 
Quite to the contrary, under the present 
economic system, the widespread deploy- 
ment of automatic factories would threaten 
jobs and undermine the financial secur- 
ity of virtually every American family. 

I claim that, if we properly 
utilized our scientific knowledge and 
our industrial capacity, we could not 
only overcome the present economic 
crisis, but we could go on to elimi- 
nate poverty altogether and guarantee 
personal financial security to every 
individual. Furthermore, this could 
be done in a manner compatible with a 
clean environment and an ecologically 
balanced world. 

The great challenge of the coming 
industrial revolution will be the 
development of an economic system 
wherein prosperity can be achieved 
without waste, affluence can be made 
compatible with the limits to growth, 
and personal freedom can be preserved 
and enhanced in a world where most 
wealth is created bv automatic 
machines. This paper is an attempt 
to formulate a plan by which this 
could be accomplished. The proposals 
contained in the following paragraphs 
might best be described as a formula 
for Peoples' Capitalism, or as a 
blue print for Jeffersonian Democracy 
in a modern technological society. 

Even the most casual observer of 
what goes on in the average factory, 
office, or construction project 
cannot help but notice that most of 
what is produced comes from machines 
and not human labor. Scholarly 
studies confirm this common sense 
observation, showing that the over- 
whelming percentage of productivity 
i her eas es over tfie pas t two hundr ed 
years have resulted from technological 
progress, not from harder work or 
longer hours. Whether we like to 
admit it or not, most of what we 
have is produced by machines, not 

This book is an attempt to address 
some of the fundamental problems of 
income distribution and capital owner- 
ship in a society where most of the goods 
and services either are, or could be, 
produced by machines rather than people. 
It questions the adequacy of conventional 
economics for the present, as well as 
for the future. It argues that the pri- 
mary cause of the recent economic crisis 
is not a lack of resources or insufficient 
wealth-producing capacity but an unrealis- 
tic view of how wealth is created and an 
outmoded system of incentives which does 
not make use of what is available to 
produce what is needed. 

Does it not then seem odd that 
more than two-thirds of our total 
output is distributed as compensation 
to labor? Might not such a large 
discrepancy between how wealth is 
created and how it is distributed 
distort the entire structure of the 
free market economy? Consider, for 
example, that distributing the bene- 
fits of two centuries of productivitj 
increases primarily through wages 
and salaries has increased labor 
costs so high that employers cannot 
afford to hire workers even when 
there are jobs which need doing. 
Thus, we have massive unemployment 
even while: 



BOX 1579, PALO ALTO CA 94302 

♦Cities need rebuilding, 

*New sources of energy need to be 

♦Pollution control needs to be 

♦Better health care delivery needs 
to be provided, 

♦The environment needs to be pro- 
tected and services of every 
kind need to be improved. 

in computers and manufacturing tech- 
nology suggest that mankind may be 
on the threshold of a new industrial 
revolution. Within two decades it 
may be practical for computer- 
controlled factories and robots to 
produce virtually unlimited quantities 
of manufactured goods, and to even 
reproduce themselves at continuously 
decreasing costs. 

Yet no one can afford to hire people 
to do these jobs. 

On the other hand the lack of any 

alternative to wages and salaries as a 

source of income creates such strong pres- 
sures for job security that: 

*Waste is encouraged, 

♦Featherbedding and restrictive 
work rules are commonplace, 

♦Pollution is condoned, 

♦Obsolescence is planned, 

*Mass advertising of trivia is 
considered necessary, 

*Unwise growth goes unchecked. 

And much of what people get paid 
for doing everyday in offices and factories 
throughout this land could be eliminated 
without affecting the production of goods 
or services whatsoever. 

Furthermore, even if make-work and 
waste could succeed in producing "full 
employment", there would still be millions 
of Americans outside the wage and salary 
income distribution channels. The em- 
ployable labor force makes up only about 
40 percent of the population. Thus, even 
though practically every business in the 
country could easily expand production, 
and would gladly do so if markets were 
available, the lack of purchasing power 
of people without jobs makes such expansion 
impossible. The result is that our 
enormous productive potential is never 
fully applied to our clear and urgent 
human needs. 

Rising Expectations vs. Declining Resources 

The two following charts show that 
long-term gains in productivity are closely 
correlated with investment. There is 
little reason to doubt that this correla- 
tion will continue into the future. In 
fact, recent technological developments 


1960-72 AVERAGES / ,• 







UJ — 

SWEDEN • •/ 


*> 6 



ITALY /• •gerhuny 

— z> 


-.,0 4 


II / • M " IM 

<S £ 




5" 2 





5 10 15 20 25 30 



1960 - 1970 

Figure IV-4. What causes a nation's productivity to grow? This chart shows 
that countries with a high rate of investment have high productivity growth, 
and vice versa. This implies that productivity growth is not serendipitous or 
beyond human control. Instead, it is the direct result of economic policies 
which promote investments in new technology and in more efficient plants 
and equipment. 


PRIVATE ECONOMY, 1950-197* ■ 20 


1950 52 54 56 58 60 62 64 66 68 70 72 74 

Figure IV-5. Productivity (i.e.. output per man-hour) is closely correlated 
with the amount of sophisticated tools and capital equipment per worker. 
The data shown here, together with that in Figure IV-4 strongly imply that 
U. S. productivity could be increased by increasing the capital investment 

Investments in such technology al- 
most certainly will produce major 
productivity gains for many decades 
to come. 



BOX 1579, PALO ALTO CA 94302 

Unfortunately, the present mech- 
anisms for increasing investment in high 
technology industries (such as tax credits 
for big business) serve to increase the 
already enormous concentration of wealth 
and power in the hands of a few (A U.S. 
Department of Commerce Survey of Current 
Business report on stockownership in the 
United States dated November 1974 states 
that one percent of the families in the 
U.S. owns over 50 percent of all stock 
and that 5 percent of the families own 
73.7 percent by value of all stock.), while 
leaving the majority of the population 
just as dependent on wages and salaries 
as ever. Clearly robot factories will be 
a direct threat to the economic security 
of almost every American unless some al- 
ternate means of investment financing 
and income distribution can be found. 

Peoples ' Capitalism is a proposal 
which addresses all of these issues simul- 
taneously. It offers a simple, straight- 
forward solution to each. Peoples 1 
Capitalism could be instituted in the 
United States without any changes in our 
constitutional form of government. In 
fact, far from altering any of the funda- 
mental principles upon which this country 
was founded, this plan would revitalize 
the free enterprise system, and realize 
the ideals of Jeffersonian Democracy in 
post-industrial America. 

Specifically three new institutions 
are proposed: 

1. A National Mutual Fund (NMF) 
is suggested to finance capital investment 
for increasing productivity in socially 
beneficial industries. The NMF would be 
a semiprivate profit-making investment 
corporation which would be authorized by 
Congress to borrow money from the Federal 
Reserve System. It would use this money 
to purchase stock from private industry 
for the modernization of plants and machin- 
ery and the introduction of advanced 
computer-based automation. Profits from 
these investments would be paid by the 
NMF to the general public in the form of 
dividends. By this means, the average 
citizen would receive income from the 
industrial sector of the economy quite 
independently of employment in factories 
and offices. Every adult citizen would 
become a capitalist in the sense of de- 
riving a substantial percentage of his or 
her income from dividends paid on invested 

2. A Demand Regulation Policy 
(DRP) would be instituted in parallel 
with the NMF in order to provide 
sufficient savings to offset NMF 
investment spending. This would 
prevent short-term demand-pull in- 
flation. The DRP would withhold 
income from consumers by mandatory 
payroll deductions and convert it 
into high-interest five-year savings 
bonds. Deductions would be graduated 
according to income (low-income 
persons would have little withheld, 
high- income more) and would be ad- 
justed monthly according to a formula 
based on the best available indicator 
for inflation. The DRP would allow 
high rates of investment and the 
accompanying high employment and high 
production while preventing excess 
demand from forcing prices upward. 

3 . A Federal Department of 
Science and Technology is also sug- 
gested to focus modern technology 
more directly on problems relevant 
to human needs. 

Peoples' Capitalism is simple 
in concept yet its implications are 
truly breathtaking. 

*It offers a solution to 
recession and inflation 

Increased availability of 
investment capital would get 
the economy moving again. 
Inflation would be controlled 
in the short run by DRP 
savings , and over the long 
run by increased productivity 
resulting from higher rates 
of investment. 

*It resolves the fundamental 
conflict between economic 
growth and environmental 

NMF dividends to individuals 
would reduce the pressures for 
jobs and growth at any cost. 
Increased efficiency in pro- 
duction would reduce waste 
and provide the resources for 
improved pollution control 
and environmental conservation. 

*It promises a degree of individ- 
ual freedom based on personal 



BOX 1579, PALO ALTO CA 94302 

financial independence which is 
unprecedented even among Utopian 
proposals . 

Every adult citizen would possess 
a personal source of independent 
income. People would be econom- 
ically free to live where they 
wished and to work at what they 

*It offers a cure to poverty and 
old age security without taxing 
the rich to give to the poor. 

NMF public dividends would be 
generated by wealth producing 
capital investments. The NMF 
would be a profit-making, income- 
producing investment corporation, 
not a tax-consuming welfare 
program . 

*It achieves economic equity with- 
out destroying incentives to 
individual excellence. 

The NMF would distribute the 
profits generated by high tech- 
nology industry to everyone. 
Persons with ambition could 
afford to develop their talents, 
and society would be able, with 
clear conscience, to reward its 
high achievers. 

*It opens up an entirely new 
road to economic development 
for emerging nations which 
completely by-passes the 
social dislocations of clas- 
sical industrialization. 

In developing nations, Peoples' 
Capitalism could finance 
automated industries and 
robot factories which would 
pay dividends to farmers and 
villagers directly. Economic 
development would be achieved 
without converting the rural 
population into industrial 
workers or concentrating them 
in congested urban areas. 

of the Robot Revolution is published 

Kensington, MD 20795 

$3 . 95 paperback 

It is distributed by: 


P.O. Box 456 

Minneapolis, Minnesota 55440 



2337 Philmont Avenue 

Huntington Valley, Pennsylvania 19006 



BOX 1579, PALO ALTO CA 94302 


by Dennis Reinhardt 

DAIR Computer Systems 
8?0 Garland Dr. 
Palo Alto, CA 9^303 
(415) 326-1534 

"... I have always discovered after the fact that, if anything, we didn»t plan big enough. 
I did not forsee the size of General Motors . .." 

- Alfred P. Sloan, Jr. 


Some of the technical problems involved 
in producing a small computer with the 
hardware capacity of the human brain 
are explored. Starting with assumptions 
differing by two orders of magnitude 
concerning the capacity of the brain 
and by 25$ for the long term growth 
rate in semiconductor chip density, 
it is hypothesized that the technology 
for manufacture of a chip or small system 
having the hardware capacity of the 
human brain can be projected for the 
period 26-48 years from now. 


The measure which is most useful 
in comparing the capacity of the human 
brain with the capacity of the computer 
is information content, measured in 
bits or bytes. As more architectural 
information emerges from brain research, 
other more useful measures might be 
suggested. Carl Sagan (1977), quoting 
data from the work of Britten and Davidson 
(1969) has placed the information content 
of the human brain at 10**13 bits(l). 
Isaac Asimov (1963) placed the figure 
at 10**15 bits(2). Using the definition 
that 8 bits is 1 byte, we obtain estimates 
for brain capacity of 1.2*10**12 to 
1.2*10**14 Bytes (abbreviated "B"). 

Since most personal computer 
systems have 4,096 to 65,536 B of high 
speed memory with 16,384 B being typical, 
we estimate the capacity difference 
between the two, to be 7.6*10**7 to 

Thus, one measure of the hardware 
gap between the human brain and a personal 
computer suggests it is impossible to 
ever achieve equality. Marilyn Ferguson, 
writing about discoveries in brain research, 
states: "A computer sophisticated enough 

to handle the functions of a single 
brain 1 s ten billion cells would more 
than cover the face of the earth. "(3) 
While such a computer would be massive, 
it is not true that the surface of the 
earth would be covered. A 3330 
type disc contains approximately 300 MB 
of data and occupies about 2+ sq. 
meters, after allowance for access 
space, at a cost in the area of $50,000. 
Thus, a 1.2*10**12 B system would need 
4000 (!) disc drives and might occupy 
8,000+ sq. meters (80,000 sq. ft.) and 
cost over $200 million. More exotic 
technologies permit cost and space require- 
ments to be reduced substantially. 

Without pursuing the issue of 
todays hardware cost any further, when 
might this kind of configuration reach 
a personal computer size, contained 
on an integrated circuit or small packaged 
system? The human brain has been evolving 
for millions of years while the computer 
has existed for only a few decades. 
Considering the enormous capacity gap 
between the two, as measured by 
the ratio of their byte storage capacity, 
what are the prospects for evolution 
of the computer over the next few decades? 


It has been the semiconductor 
industry which has paced the state of 
the art in computer technology with 
the development of increasingly complex 
circuits on a single integrated circuit 
chip. The chairman of Intel Corp. , 
Robert N. Noyce, has said that "... 
the number of components per circuit 
in the most advanced integrated circuits 
has doubled every year since 1959, when 
the planar transistor was developed." 
If we were to extrapolate this rate 
of growth forward in time, we would 
have a low estimate of when "brain" 



BOX 1579, PALO ALTO CA 94302 

hardware could be produced. To do this 
we solve: ' 

2**N = 7.6*10**7 

for N to obtain a value for N of 26 
years. Thus, if the miniaturization 
trends of the last 18 years continue 
for another 26 years, semiconductor 
technology could be at the point where 
personal computers with a hardware 
capacity equal to the human brain in 
terms of number of bits could be produced. 

Turning to another source, the 
chairman of Texas Instruments, Mark 
Shepherd, Jr. , has presented data which 
exhibits a growth rate closer to 60# 
per year rather than 100# (5) and extrapolates 
his data forward for the next 20 years. 
If we use the 60# growth figure and 
the higher value for the ratio of brain 
capacity to today* s personal computer, 
we can project a high figure by solving: 

1.6**S - 7.6*10**9 

for S, to obtain a high value of 48 
years. In other words, we obtain a 
range of estimates for the brain computer 
of 26-A8 years, depending upon growth 
rates in semiconductor complexity and 
brain capacity used for the calculation. 

This range estimate is also subject 
to the caveat that the growth in complexity 
might slow or halt altogether, depending 
upon technical problems relating to 
ultimate physical limits, market resistance 
to processor chips which are software 
incompatible with previous generations, 
or escalating cost of the new production 
facilities. Also, neither of our quoted 
sources extrapolated the data as far 
as we have. 

Determination of the required 
word size is another hardware problem 
which would need to be solved. The expression 
for the word size using our 1.2*10**12 
Byte brain size is seen to be 
ln(1.2*10**12)/ln(2) = 40.1 bits. Evaluating 
the larger brain size yields a figure 
of 46.8 bits. Thus, about 40 to 47 bits 
minimum are needed to duplicate the 
addressing within the human brain. 
The implication for the brain computer 
is that it have a 64 bit, or similar 
instruction word length since there 
would be extra bits required in the 
instruction word for opcode, address 
mode, etc. At present, the standard 

for the personal computer market 

is the 8 bit processor. The 16 bit 

processor probably has no more than 

10# of the market. Shepherd had predicted 

the 128 KB 32 bit microprocessor for 

the mid 1980* s. If such processors become 

available at that time, it is likely 

that boards or board sets with that capability 

could be introduced sooner since the 

microprocessor chip can be developed 

independently of the memory. 


If indeed the hardware problems 
can be solved, how would the software 
for a computer of this capacity be developed? 
Using standard estimating techniques 
and assuming that those 1.2*10**12 B 
in the "brain" computer are produced 
by programmers working at an average 
productivity of 100 - 1000 instructions 
(64 bit word) per month (6), we find 
that between 12 million and 120 million 
man-years of programmer effort are required 
By way of comparison, the emergence 
of man on this planet is thought to 
have occurred less than 10 million years 
ago. Even worse, other studies suggest 
that programming effort rises exponentially 
with the size of the programming problem 
with an exponent of 1.5 (7). 

We are forced to conclude that a 
"brain" computer will never be 
programmed in this fashion. Given this 
conclusion, what relaxation of assumptions 
makes it conceivable? Certainly, one 
way it could be done is if we relax 
the assumption that all 1.2*10**12 B 
have to be programmed in by a programmer. 
For example, if the programmer were able 
to specify the first 3*10**6 B and the 
computer learned the rest on its own, then 
we have reduced the task several orders 
of magnitude. Probably none of us knows 
the number of bytes required for a computer 
to be able to organize almost all of the 
information coming into it without programmer 
modification over a period of years. None 
of us will know until someone does it. 

Even scaling the problem down to 
3*10**6 B gives software development time 
estimates between 31 and 310 man-years. 
At present, continuous speech recognition 
systems require 4.5*10**5 to 9*10**5 B. (9) 
and are running only on fairly large computers 
such as the PDP-10. An overall estimate of 
3*10**6 B implies that continuous speech 
recognition is about 10+£ of the entire 



BOX 1579. PALO ALTO CA 94302 

job in a brain computer, 
one knows. 

But again, no 

It might be argued that we have 
unnecessarily concentrated on programmer 
entry of algorithms when the bulk of the 
brain information might be made up of the 
"data base". This might be the crucial 
distinction between our view of automated 
intelligence and the state of the art in 
present software systems. In present 
practice, the "data base" is operated upon 
statically by the computer system. It 
makes no difference to the algorithms over 
time what the range of data entered is, how 
often it is entered, or its source. One 
unchanging computation fits all cases until 
the programming staff makes another release. 
We forsee an intelligent system modifying 
its behavior based upon the data it ex- 
periences. This means that the fundamental 
algorithms used in processing must be 
capable of evolution under the control of 
the computer itself. Part of the evolution 
can be controlled by the software manufacturer, 
but much of it must be local to the machine 
and its owner in order that the computer 
remain intelligent within its surroundings. 

Within the "hobby" subculture of the 
personal computer market, software development 
takes place outside the Confines of an indi- 
vidual's organizational affiliation. Often 
there is more creativity exhibited on one»s 
own work than in the work done for an 
employer. Furthermore, there seem to exist 
methods with lag times of 3-6 months for 
transmitting software through most of the 
hobby market. With an estimated 50,000+ 
systems sold and somewhat less actually 
operating, there exists a large pool of 
people who could participate in developing 
a "brain" computer. It is already considered 
true by some in the personal computer 
industry that it leads other, more established 
segments of the computer industry in terms of 
accomplished innovation. It seems likely 
that progress will be made in the process of 
automating the intelligence acquired by man 
and that significant contributions can 
potentially come from the hobbyists in the 
personal computer market. Already some 
products have been introduced which automate 
some human like functions with modest memory 
requirements of 8,000-48,000 B. Examples 
are isolated word discrimination (9) and 
computer controlled speech (10). As this 
industry comes to understand how to im- 
plement functions concisely, the enormous 
software cost estimates for developing a 
computer brain can be lowered closer to 


Having considered some of the hardware 
and software development problems to be 
solved in developing the brain computer, 
let us raise an issue suggested by this 
investigation: if only the first 3*10**6 B 
of the computer are programmed and the 
remaining 1.2*10**9 B learned and organized 
by the machine, then in what sense are 
the capabilities of the machine "pre- 
destined". Well over 99# of the machine *s 
information has entered as a result of 
interaction with the environment. It might 
well be that the only practical way to 
develop a brain computer is for it to 
be self determining. 


A personal computer with hardware 
capacity comparable to the human brain 
might be possible 26-48 years from now, 
and its near term ancestor, the 32 bit 
microcomputer can be forseen within the next 
few years, before the end of the 1980* s. 
The software does not yet exist and it is 
possible that the cost of producing it 
might be prohibitive unless we are able 
to implement innovative approaches to a 
programming effort of this magnitude. 


(1) THE DRAGONS OF EDEN: Speculations on 
the Evolution of Human Intelligence. Carl 
Sagan. Random House, Inc., 1977. 

(2) THE HUMAN BRAIN: Its Capacities and 
Functions. Isaac Asimov. Mentor Books, 

(3)THE BRAIN REVOLUTION. Marilyn Ferguson. 
Bantum Books, Inc., 1973. 

(4) "Microelectronics". Robert N. Noyce 
in SCIENTIFIC AMERICAN, Vol. 237, No. 3, 
pages 63-69; September, 1977. 

(5) "Distributed Computing Power: a Key 
to Productivity". Mark Shepherd, Jr. in 
COMPUTER, Vol. 10, No. 11, pages 66-74; 
November, 1977. 

(6) "The Cost of Developing Large-scale 
Software". Ray W. Woverton in IEEE 
pages 6I5-636; June, 1974. 



BOX 1579, PALO ALTO CA 94302 

Brooks, Jr. Addis on-Wes ley, Inc., 1975. 

(3) "An Assessment of the Technology 
of Automatic Speech Recognition for 
Military Applications:. Bruno Beek 
Edward P. Neuberg and David C. Hodge in 
pages 310-322; August, 1977. * 

(9) "Introducing SPEECHLAB - The First 
Vocal Interface for a Computer". Horace 
Snea and John Reykjalin in POPULAR 
ELECTRONICS, Vol. 11 No. 5, pages ^3- 
50; May, 1977. 

(10) "Friends, Humans and Countryrobots 
Lend Me Your Ears". D. Lloyd Rice in BYTE, 
issue 12, pages 16-24; August, 1976. 



James S. Albus 

New World Books 

4515 Saul Road 

Kensington, Maryland 20795 


The inventor of the neurophysiologi- 
cal model Cerebellar Model Arithmetic 
Computer (CMAC) which won in the 1976 
Industrial Research Magazine IR-100 com- 
petition describes his work in brain mod- 
eling and robot control. CMAC demonstra- 
tes the capacity to learn, generalize, 
recognize; patterns, perform associative 
recall, compute multivarient analog 
functions and decompose input commands 
into sequences of output commands in a 
context sensitive manner. Methods of 
implementing this model on a microproces- 
sor are discussed. 

Evidence is given that clusters of 
neurons with such properties are arranged 
in hierarchies in human brains so as to 
produce AND/OR task decompositions. At 
the lowest levels in the motor system 
these clusters transform coordinates and 
compute servo functions. At middle levels 
they decompose input commands into se- 
quences of output commands which give rise 
to goal directed behavior patterns. Mech- 
anisms by which feedback can alter these 
decompositions to compensate for pertur- 
bations and uncertainties in the environ- 
ment are described. At the highest levels 
of the hierarchy are goal selecting and 
evaluating mechanisms which are used for 
planning and problem solving. 

The possibility of implementing such 
a hierarchy on a network of hobby comput- 
ers is discussed. 

Computation by Table Look-up 

CMAC is a computing device which 
accepts input variables and produces an 
output which is some function of the in- 
put variables. There may be up to twelve 
input variables which can be either con- 
tinuous or binary in any combination. 
CMAC's internal operations are entirely 
digital and its output is a digital 
number which may, of course, be converted 
to an analog voltage. 



CMAC computes by transforming 
each input variable into a set of 
intermediate variables which are 
combined to select a set of weights. 
These weights are then summed to 
produce an output. The CMAC output 
is thus a function of the input 
variables , and the value of the 
output for every possible configu- 
ration of the inputs is determined 
by the weights. 

The computation of mathematical 
functions by table look-up is, of 
course, a commonly used technique 
for functions of from one to three 
variables. For example, trigonomet- 
ric functions of one variable are 
often looked up in tables rather 
than computed by numerical methods. 
Particularly when combined with 
interpolation techniques, table 
look-up is a powerful tool. 

For functions of four or more 
variables, however, conventional 
table look-up becomes impractical. 

If each variable can take on Rr dis 

tinguishable values the number of 
table entries required for N vari- 
ables is R . Thus, even if each 
input variable is limited to as few 
as 16 values over its range, a table 
for storing a four dimensional func- 
tion would require 16 or 65,536 
entries. Clearly, table look-up is 
impractical for most functions 
involving more than four variables. 

CMAC, however, does not require a 
unique table entry for each possible 
input vector. Instead, CMAC maps 
each input vector into a multipli- 
city of table entries. This is il- 
lustrated in Figure 1. The value 
of the CMAC output is equal to the 
arithmetic sum of the contents of 
the selected memory locations. Now 
the number of ways a set of "a" 
elements can be selected from a 
table with "b" entries is the number 
of combinations of "b" things taken 
"a" at a time. This, in practical 
cases, is much much larger than "b". 

BOX 1579. PALO ALTO CA 94302 

Thus CMAC can utilize a relatively few 
memory locations to represent a space de- 
fined by N input variables. 

Of course, the fact that each possi- 
ble CMAC input vector selects a unique 
set of memory locations rather than a 
unique single location implies that any 
particular location may be selected by 
more than one input vector. CMAC turns 
this necessity into a convenience by an 
algorithm which operates such that any two 
input vectors which are similar (i.e.. 
close together in input space) map into 
highly overlapping sets of memory loca- 
tions as shown in Figure 2. 

This gives CMAC the property of 
generalization, i.e., CMAC tends to pro- 
duce similar outputs for similar inputs. 
In Figure 2, input vector S> 2 selects 
three out of four of the same memory loca- 
tions as Si- Thus, the output h(S2) will 
be similar to h(Sx) differing only by 
the contents of the single location which 
is not in common. 

The amount of overlap between sets 
of selected memory locations is controlled 
such that as the input space distance 
between two input vectors increases the 
amount of overlap decreases. Finally, at 
some distance the overlap becomes zero 
and the sets of selected memory locations 
are disjoint. At that point input S 2 can 
be said to be outside the neighborhood 
of generalization of Sx . The value of 
the output h(S2) is thus independent of 
h(Si) . 

An example of a neighborhood of 
generalization of CMAC can be seen in 
Figure 3. In this example, the input 
vector consists of two input variables 
s^ and S2 with range 0<sx^360 and 
0-cS2<:180. There is unity resolution along 
each" 1 variable axis so that R is 360 for 
S-. and 180 for s 2 . In Figure 3. 32 
weights are selected for each input vector, 
and thus any two input vectors which dif- 
fer by only one resolution elements will 
have 31 weights in common. Not until two 
inputs are at least 32 resolution elements 
apart do they map into disjoint sets of 
weights. If the weights are initially all 
zero, the value 1 can be stored at some 
point such as S x = (90,90) by placing 
1/32 in each of the 32 weights selected by 
St. Following this operation one will 
find that a second input vector S2=(91,90) 
will produce the output 31/32. This is 
because S2 shares 31 weights with the 
vector Si. A third vector S3=(92,90) or 

t s 




(S 4 = (90.92)) will liiivo an output 
30732 because oi sharing ISO weig 
with Sj_ . etc. The result is t ha 
the CMAC memory generalizes. Ad 
jacent memory locations are not 
dependent, and a plot ot values 
stored at each point in input sp 
shows the appearance oi a stretc 
rubber sheet. Pulling owe point 
a particular value, as in Figure 
affects adjacent points. 

Generalization has the distinct 
advantage that training- (or data 
storage) is not required at every 
point in the input space. Training 
on a suitably representative subset 
of input vectors is adequate. For 
example, if one wishes to store the 
function sin(s 1 ) sin(s 2 ) to a resol- 
ution of one degree in a convent iona 
memory, data would have to be stored 
at each of the 360X180 (more than 
64.000) possible inputs. In Figure 
4. data was entered at only 175 
input points scattered over the 
input space and the tendency of the 
memory to generalize fills in the 

Of course, generalization is 
disadvantageous if radically dif- 
ferent outputs are required for 
highly similar inputs. In most 
control problems however, such dis- 
continuities do not obtain. Indeed 
most control functions have simple 
S-shaped characteristics along each 
variable axis. The complexity of 
control computation in multivarient 
systems typically derives from cross- 
products which effect the slope of 
the function, or produce skewness . 
or non-symetrical hills and valleys 
in various corners of the N dimen- 
sional space. As can be seen from 
the function shown in Figure 6 these 
are the type of functions which 
CMAC can readily store and hence 

A Neurophysiological Model 

A neurophysiological theory 
upon which CMAC is based was devel- 
oped independently in England by 
David Marr and in America by the 
author. It a/as published first by 
Marr in 1969\ The cerebellum had 
for some years been known as one 
of the areas in the brain respon- 
sible for controlling the limbs, 
hands, eyes and fingers in rapid. 



BOX 1579, PALO ALTO CA 94302 

precise, coordinated movements. It was 
known to receive neural input from motor 
areas of the cerebral cortex as well as 
feedback signals from the muscles, joints 
tendons and skin. Marr and myself inde- 
pendently combined this knowledge with 
new information from an elegant series c :' 
experiments by Eccles. Ito and Szenta- 
gothai^ into a theory of how these input 
signals act through the various cellular 
interconnections to produce outputs with 
appropriate values. The result was a 
theory of how conditioned reflexes could 
be stored and recalled (i.e., computed). 
This theory has rapidly become one of the 
most widely accepted working hypotheses 
among cerebellar neurophysiologists . 

The CMAC formalism not only gives 
mathematical structure to the original 
Marr-Albus theory but suggests analytic 
procedures by which electronic circuits 
with similar properties can be synthesized. 

For example. CMAC first transforms 
the single precise value of each input 
variable into a multiplicity of less 
precise values on a set of intermediate 
variables. This is analogous to the 
function accomplished by sensory end 
organs in biological systems. in the 
body, the angular position of a joint, 
the tension in a tendon, the velocity of 
contraction of a muscle, are all precise 
physical parameters each of which is en- 
coded by a multiplicity of sensory organs 
into firing rates on neuron axons which 
are relatively imprecise information chan- 
nels. CMAC emulates this in the S-*M 
mapping by which the value of each vari- 
able in the input vector S = (sj_,S2 ,S3, . . . 
sn) is transformed into a set of inter- 
mediate variables. 

Figure 5 illustrates the essential 
characteristics of the S* M mapping. In 
this illustration the two input variables 
s^ and S2 are represented with unity 
resolution on the range to 16. The 
range of each input variable is also 
covered by four intermediate variables 
of lower resolution. 

In Figure 5, s^ is mapped into a set 
mi composed of four intermediate variables. 


'1 * 2 ' 3 ' 4 " 


Cx = (A, B, C, D, E) 

C 2 = {F, G, H, J, K) 

C3 - -M, N, P, Q, R- 

C4 = -S, T, V, W, X- 

For every value of sj_, there 
exists a unique set of elements m^* 
one from each set of intermediate 
variables in m^ . such that the value 
of Sj uniquely defines the set mi* , 
and vice versa. For example, in 
Figure 5 the value s-^7 maps into 
the set m 1 *--B. H, P. V- and vice 
versa. Similarily. the value 32=10 
maps into the set ni2*= - c. j. q. v. . 
and vice versa. 

In the cerebellum, incoming 
sensory neurons enter the granular 
layer where they make contact with 
granule cells. A system of negative 
feedback regulates the overall 
activity of the granular layer so 
that a small and relatively fixed 
percentage of the granule cells are 
allowed to become active. This is 
stimulated in CMAC by the combina- 
tion of intermediate variables to 
select a set of weights . For the 
example in Figure 5, the set m-,* = 

B, H, P. V and m 2 * = c , j, q, v. 
combine to select the" set of weights 
-Be. Hj , Pq, Vv- . 

Finally, in the cerebellum 
the Purkinje cell sums the influ- 
ence of the active granule cells 
through a set of weighted synaptic 
connections. Similarly in Figure 5 
CMAC sums the selected weights. 














Thus, the input S=(7, 10) 
produces the output h(S)=4. The 
particular set of weigh~ts shown in 
Figure 5 defines the function in 
Figure 6. 

At every point in input space, 
four weights are selected whose sum 
is the value of the output. As the 
input vector moves from one point 
in input space to an adjacent point, 
one weight drops out to be replaced 
by another. The difference in 
value of the new weight minus the 
old is the difference in the value 
of the output at the two adjacent 
points. Thus, the difference in 
adjacent weights is the partial 
derivative (really the partial dif- 
ference) of the function at that 
point. For example, in Figure 5, 
if the input vector moves from 



BOX 1579, PALO ALTO CA 94302 

S-(7. 10) to S (8. 10) the weight Be 1.0 
drops out and is replaced by Cc=2 t 0. The 
value oi' the output thus changes from 4 
to 5. 

CMAC provides a mathematical t'or- 

moliam whi/ili ic citnnlo vdt nrppi<!P iinri 

which accurately reflects the functional 
properties of a cerebellar output cell 
and its associated interneuron network. 
Furthermore, the CMAC formalism is suf- 
ficiently general that it can be said to 
closely approximate the functional proper- 
ties of output neurons and their associated 
interneuron nets in a large number of cor- 
tical regions and subcortical nuclei. A 
multiplicity of CMAC's can functionally 
simulate large sections of cortex and 
entire processing nuclei. CMAC thus pro- 
vides a mathematical tool which may be 
applied to analysis of sensory-motor sys- 
tmes in many different areas of the brain. 

Input to CMAC is a vector which 
defines an N dimensional space. At any 
instant of time the input vector defines 
a point in input space. If any one of the 
input variables (component of the vector) 
is time dependent then as time progresses 
the input vector moves through space de- 
fining a trajectory. We can now speak of 
an input trajectory through N dimensional 
input space. At any instant of time CMAC 
accepts an input vector and produces an 
output which is a scalar. A multiplicity 
of CMAC's produces an output vector. 

Thus, for a cell nucleus represented 
by a multiplicity of CMAC's. at any in- 
stant of time there exists a one-to-one 
mapping from each input point (vector) in 
input space into an output point (vector) 
in output space. And correspondingly for 
every input trajectory there is an output 

Let us now explore how these basic 
CMAC concepts can be applied to the sub- 
ject of higher level behavior such as is 
required in the performance of a behavioral 
sequence, or task. 

A Sensory-Motor Hierarchy 

Simply stated, any task can be 
broken down into subtasks in a hierarchical 
way until finally at the bottom of the 
hierarchy there are action primatives such 
as motor neuron firing rates. This hier- 
archical decomposition of complex tasks, 
or behavior, is not new having been sug- 
gested many times by workers in behavioral 
psychology, linguistics and various other 
fields such as military command and con- 
trol, and manufacturing engineering. The 

entire cor.cept of mass production 
is based on the principle of break- 
ing down the task of making a pro- 
duct into a series oi elemental 
operations which are so simple that 
they can be taught to low-skilled 
workers in a short period oi time. 

Task decomposition immediately 
suggests a control hierarchy where 
each level in the hierarchy accepts 
input commands (or tasks) from the 
next h.igner level and responds by 
issuing ordered sequences oi output 
commands (or subtasks) to the next 
lower level. In extremelv simple 
cases these sub task generators may 
merely produce sequences of pre- 
recorded outputs. But m more com- 
plex systems sensory teedback from 
the environment, or from the 
system being controlled, may alter 
the output sequences in one way or 
another. Sensory feedback may 
consist of interlock signals to 
provide sequential tuning, or may 
incorporate any number oi analog or 
digital variables which modify the 
transier function of the subtask 
generator . 

Suppose, for example, that 
the goal is to program an industrial 
robot to assemble a gasoline engine 
This task can be broken down into 
sequences of simpler tasks such as 
fetch parts a and b. insert a into 
b while nulling off-axis forces, 
fasten part c to assembly ab and 
so on. Each of these tasks can be 
broken down further into sequences 
of elemental movements such as 
reach, grasp, follow specified 
trajectory, etc. These elemental 
movements can themselves be broken 
down into sequences of positions in 
physical space. Finally, each 
point along the physical trajectory 
can be transformed into a coordinae 
system defined by the physical 
structure of the robot and its 
actuators. This concept is illus- 
trated in Figure 7. 

At each level in such a 
hierarchy there are two types of 
input. First there are input com- 
mands from a higher level. At the 
very top these may come from the 
shop foreman or from a production 
planning and scheduling program. 
At all other levels, commands 
originate as outputs from higher 
levels in the hierarchy. 



BOX 1579. PALO ALTO CA 94302 

Each hierarchical level also 
receives feedback signals which report the 
position and motion of joints or convey 
information from sensors monitoring force, 
vouch, or visual position of objects in 
'.he environment. In some cases this feed- 
oack is used for timing purposes in order 
to coordinate sequences of actions with 
conditions in the environment. Feedback 
may also indicate the recognition of pat- 
terns of events, or shapes and locations 
of objects iv the environment, or even the 
movement of coordinate systems and frames 
of references. Particularly at the higher 
levels in the hierarchy, feedback is 
highly processed through many layers of 
an ascending hierarchy of CMAC like pro- 
cessing nuclei . 

In biological systems as well 
there may be a sensory-data processing 
hierarchy, which runs parallel to and in 
the opposite direction from the motor- 
vehavior generating hierarchy. This pro- 
cessing hierarchy receives input at the 
lowest level directly from sensory trans- 
ducers. At each level input vectors (and 
trajectories) are transformed into output 
vectors (and trajectories). These out- 
puts are then passed on to the next higher 
level as inputs. The purpose of each 
processing module is to transform input 
vectors into output vectors which are 
optimally configured to serve as feedback 
to the motor-generating hierarchy at that 
level. We can therefore hypothesize a 
feedback link from each level in the 
sensory-processing hierarchy to corres- 
ponding levels in the motor-generating 

Furthermore the efficiency of the 
sensory-processing hierarchy is enormously 
enhanced if there are complementary links 
from the motor-generating hierarchy to 
the sensory-processing hierarchy. This 
type of information pathway is what neuro- 
psychologists call an "efference copy." 
This information tells the sensory pro- 
cessing hierarchy what the body is doing 
so that, among other things, it can dis- 
tinguish sensory data resulting from 
movement of the body from sensory signals 
resulting from movement of objects in the 
environment of the eyes and a rotation of 
the room about the eyes. More generally, 
it enables the processing system to per- 
form context sensitive filtering, and 
indeed, to do predictive filtering. 

It is, of course, possible to 
turn off the lowest level of the motor 
hierarchy without disabling the entire 
processing-generating hierarchy. When 
in this mode, the generating hierarchy 

can be used to produce signals which 
facilitate the operation of the 
sensory-processing hierarchy. Activ- 
ity in the generating hierarchy may 
now be better characterized as hypo- 
theses, instead of tasks and sub- 
tasks. Sensory input from the 
environment is now analysed in con- 
junction with hypotheses from the 
generating hierarchy. If the hypo- 
theses are correct , they will assist 
in sensory recognition. If only 
nearly correct they can be "pulled" 
by feedback from the processing 
hierarchy. When a particular hypo- 
thesis is successful in generating 
predictions which match incoming 
sensory data the entire processing- 
generating hierarchy "locks on" to 
the incoming sensory data. This gives 
the hierarchy the ability to recognize 
and track lengthy sequences of input 
signals even in the presence of noise 
and interference from similar signals. 
This lock-on phenomenon at several 
levels gives the hierarchy the 
ability to recognize phrases and 
patterns with several different, but 
harmonious, periodicities. For 
example, the affinity of the ear for 
rhythmic patterns of music and 
poetry may arise from synchrony in 
hierarchical looping structure such 
as shown in Figure 8 where each of 
many different loops locks-on to 
rhythmic patterns at its own level. 
The cross-coupling in the sensory- 
motor hierarchy suggests a mechanism 
for the strong tendency of people 
to dance, or tap their feet in time 
with rhythmic sounds. The muTti- 
plicity of levels in the hierarchy 
suggests a mechanism for simultan- 
eously locking-on to many different 
levels of rhythm and "meaning" in 
both speech and music. 

Belief and Understanding 

One may hypothesize that in humans 
the functional relationships stored 
in upper levels of the cross-coupled 
hierarchy of sensory-processing 
behavior-generating modules makes up 
what might best be called a "belief 
structure" which gives rise to goal- 
directed behavior. By definition the 
belief structure is isolated from 
direct contact with the motor-output 
sensory-input levels. This enables 
the belief structure to decouple it- 
self from outward manifestations of 
behavior as well as from direct 
physical sensations of the external 
environment. It is thus free to 



BOX 1579. PALO ALTO CA 94302 

generate hypotheses, and imagine the con- 
sequences. It can send hypothetical goals 
to the mid-levels of the generating 
hierarchy which cycle through an imagined 
task, thereby stimulating the sensory- 
processing hierarchy into producing a 
facsimile of expected (or remembered) 
sensory experiences. These self -induced 
sensory feedback signals generate emotional 
reactions good, bad or neutral and these 
steer the highest level goal selecting 
mechanisms toward command vectors which 
produce rewarding emotional feedback. 

Thus , a person selects goals and 
plans actions on the basis of what is 
stored in the transfer functions of his 
or her belief structure, i.e., the mid 
and upper-levels of the processing- 
generating hierarchies. 

Similarly, a person interprets 
sensory experiences according to what is 
stored in the belief structure. People 
tend to see and hear what they expect to 
see and hear. They tend to interpret 
unfamiliar sensory input as "just noise" 
or without "meaning." Sensory experiences 
which correlate with patterns stored in 
the belief structure are reassuring and 
comforting. The world becomes predict- 
able and presents few surprises, i.e., 
we "understand" and events "have meaning." 
Unexpected twists in familiar patterns 
elicit surprise, and systematic deviations 
from expected trajectories may cause 
relearning and the establishment of new 
trajectories of expectation. Sensory 
experience which does not correlate with 
stored patterns is disturbing and is 
rejected or avoided. If such stimuli 
cannot be avoided and if it is too un- 
familiar or unpredictable to be reinter- 
preted through learning, then it may 
produce emotional distress, or neuroses. 

A processing-generating hierarchy 
of CMAC modules thus provides a simple 
unifying structure which can be applied 
to a wide range of sensory-interactive 
goal-directed behavior including the 
selection of goals, the planning of actions 
the imagining of results, the learning 
of motor skills, recognition of sensory 
patterns, and the construction of habits 
and beliefs. The fact that the basic 
building blocks for such hierarchies can 
be fabricated inexpensively with presently 
available technology suggests that it may 
now be reasonable to seriously address the 
feasibility of constructing machines with 
these properties . 

Each CMAC module is an independ- 
ent computing device. It accepts 
input variables, performs its compu- 
tations, and transmits its output. 
Each module is trained, or program- 
med, to compute a relatively simple 
function of many variables. Training 
begins at the lowest levels first and 
must be well underway at each level 
in the hierarchy before it can begin 
at the next higher level. Just as 
in a child the abilities to deal with 
abstractions do not develop until 
after visual processing and motor 
coordination is accomplished, and the 
ability to read does not emerge until 
speech generating and recognizing 
have been developed, so in a CMAC 
hierarchy training must be well 
advanced at each level before con- 
sistent and clearly defined trajec- 
tories emerge to be used as feedback 
for the next higher level. 

Networks of CMAC modules compute 
in parallel and control is distri- 
buted throughout the network in a 
manner very similar to the neural 
networks of the brain. The addition 
of new computing elements at each 
level serves to refine the precision 
of the computed output at that level. 
The addition of new levels in the 
hierarchy increases the complexity 
of the behavior which can be gener- 
ated and the sophistication of the 
sensory processing which can be 

These properties suggest that it 
may be possible to implement a CMAC 
hierarchy for controlling a robot 
on a network of microcomputers , and 
to systematically increase the 
sensory-motor capacity by incre- 
mentally adding additional modules 
to the network. If the model sug- 
gested here is correct, the result 
may be that intelligence will evolve 
in a silicon and copper network in 
much the same way as it did in bio- 
logical brains. 



BOX 1579. PALO ALTO CA 94302 



tMarr, D. , "A Theory of Cerebellar 
Cortex," J. Physical, London, 202 , 
1969, 437-470 

Albus, J., "A Theory of Cerebellar 
Function," Math. Biosciences, 10 , 
1971, 25-61 

. "A New Approach to Manipu- 
lator Control: The Cerebellar Model 
Articulation Controller (CMAC)," 
Journal of Dynamic Systems, Measure- 
ment and Control, September 1975, 

. "Data Storage in the Cere- 

bellar Model Articulation Controller 
(CMAC)," Journal of Dynamic Systems, 
Measurement and Control, September 
1975, 228-233 

.Eccles, J.C., M. Ito, and J. 
Szentagothai, "The Cerebellum as a 
Neuronal Machine," Springer, Berlin, 

It is impossible in such a brief 
paper to properly acknowledge the 
source of all the ideas presented. 
A much more complete bibliography 
is contained in each of the above 
sources . 
















BOX 1579, PALO ALTO CA 94302 












*» <» ** t » r ^ 

e * 


.f * "* * £ ^ ~ 2 



BOX 1579, PALO ALTO CA 94302 


M i * e y i i r e r 
920 r-pnnis ~rive, Falc Alto, CA ?4303 

iiisti S£ ^ 









/aricus asoactr: of trio 

e discussed. The ^npha^i 

much to expose the desi 

e motivation behind var 

3 3 C t S . The pfSn^lp"! 1 t 

be that th-? ^C^-T eiubodi 
'.saxiraize the accession 
mmunication techniques to 
ssiole segment of t 
bbyist community. On the 
oole having enhanced p 
r 1 w a r e or software can 
van tag? of their advanc 
en they use the use POT 
r their communication. 

?CN£T lesion 
s h*re is not. 
gn as to show 
ions of it^ 
heme is shovn 
es an attemot 
i ty of useful 

the broadest 
he computer 

other hand, 

OHt?r in their 

ta*e good 

ed facilities 

T techniques 



[*'e r. t 

0?ve Jaulxins organized 
r o.Tiir.ittea shortly after the First 
Coast Computer raire last year; its goal 
was to tuiid a personal computer network 
(whence its nan:e). s ftar a few n-onthr 
of lively debate, it produced a design, 
wnich v?a? subsequently implemented f iv 
test purposes. ?h° design was then 
disassembled an 1 reconstructed in such a 
wav ^»s to orofit from the experience of 
the first iteration. The second 
veneration d^si^n is no* quite stab! 3 
and is beginning to generate some second 
generation test results. at this 
juncture, then, it could perhaps b» 
instructive to consider various aspects 
of the design and see just vhy they 
turned out the way they did, 

2. rae._£tourid_liilp£ 

The design of the PCNET Proceeded *rorc a 
lumber of ground rules. 

3. The network should oe cheap and 
reliable. Then Deople would he 1 ikelv 
to freely share larce files and large 
nunbers of substantial programs. If 
they nave confidence in network 
communication, they will be tempted to 
use it casually to spread the word of a 
good program, an interesting new idea or 
a reliable merchant. 

i nt 
a dm 

The net wo 
h as lea 
outers. T 

s e or a i<n 

it can beg 
t icipants 
endi ture. 

1 1 Lkely he 
its partici 

rk should own no 


sed lines cr 


hen, the network 

can coma 

without any nece 

ssity to 

inister substanti 

al funds, 

in operation with 

too few 

to justify 

a large 

Also, w i *■ h the 


highly iecentra 

iized, it 

resDcnsive to the -needs 

:> a n t s . 

?. Permanent service nodes will not oe 
needed but will be accommodated if 
oresent. 411 tne above arguments for 
decentralization also apply to this 
ground rule. However, permanent service 
nodes, such as the CTK's I've described 
elsewhere, can eliminate very strong 
incentives for specialized hardware that 
will enable nodes to automatical ly 
handle incoming traffic. 

1. There should be no geographical 
restrictions. For example, the PCHr'T 
should be suitable for communication 
across national boundaries. 

5. The PC^f-T snould accommodate remote 
/irtual terminals. This will allow 
oeople to test a program without having 
gone to the trouble of importing it and 
adapting it to a new home. This 
feature, whil*> not yet explicitly 
includes, has shaped some p^rts of the 



BOX 1579, PALO ALTO CA 94302 

*• uSKSiS-fll.ltlSliaCtlbo 

*. 22mi_2Lli£i_2asia[i_Is2uss 

The PC.^tT specification uses a number The f 
of levels of abstraction to describe the block: is 
transmission of mail and fiJes between between b 
net-work noias. Two network nodes will that the 
tyoically be connected rv a phone line, applied t 
Thpy use models tc s^nd digital bits in put onto 
*he form of t.h- au :, io tones suitable for to great 
teleohona circuits. These bits serve in the error 
their turn to serially transmit binary leads to 
3ata in 8-bit bytes, (tfodes usinrr 5-bit woul* be 
serializers or that for other reasons 
cannot directly transmit 9-bit bytes use 
a radix-4l scheme to represent binary 
data with a restricted set of 41 
characters.) These bytes are organizad 
into transmission blocks containing 
*»rror correction information sufficient 
to provide transmission (at very low 
*rror rates) for several streams of 
bytes. One stream is devoted to 
information used to control the 
Tonmunication at this level and the 
previous ones? the other streams 
transmit messases between the server 
levels of the nodes. Finally/ these 
«essag*s transmit nail and files between 
two nodes, with forwarding automatically 
Provided wnen appropriate. 





c a co 

the li 

ly re 


a si 


and r 
rap] «t 
ed ot 

a link 
nt of 

a block 
This do 
the eff 


transrai ssior 

the choice 

in the sense 

ormation is 

before it is 

es not seem 

ectiveness of 

tion/ and it 

entation than 

Noise on the phone line can cause the 
receiving 'JAKT to lose synchronization 
with the byte frames, 
completely turning the received data 
into garbage until synchronization is 
again established. We assume that the 
error detection information in a 
transmission block will have no trouble 
detecting any such garbage. Tn an 
attempt to limit the damage to single 
transmission blocks, they are separated 
by characters chosen for their utility 
in helping the receiving UART come back 
into synchronization with the byte 

af th 
in th 



the i 




3ff ec 





ese level 
ously sepa 

may be 
rbance to 
is modular 
e way the 

ver on tra 

levels of 

nterface b 

1 ine m 

sion to a 

t on the o 
, the mod 
x or full 
mission bl 

s of speci 
rated, so 
changed with 
the others. 
ity has 3lre 
choice betw 
binary ha 
nsmission a 
etween the m 
ay or way 
) phone ca 
ther levels. 
em (and UART 
duplex, a 
is tha* 


in ano 

s separ 
the PC 
age of 
e, the 
ed by ra 
the del 
be adde 
as a vir 
ther nod 

a t i o n 
NaT t 
dio li 
ays th 
d, so 
tua) t 

o grac 

hone 1 
nks or 
at are 
ite c 

fication are 

that any of 

little or no 

An example 
a ly been seen 
een radix-41 
s no effect 
t higher or 
odem and the 
not include 

answer (or 
lis with no 

By the same 
) can be half 
nd the only 
handling link 

n levels will 
efully take 
pments. For 
inks can be 
even by links 
inevitable in 
actions can 
one node can 
to programs 

Ke had originally assumed that the 
boundaries of the link transmission 
blocks and the stream blocks would be 
entirely independent; we soon 
discovered, however, that such 
independence could require a general, 
coroutine capability in the 
implementation. In order to maximize 
the 3CcessibiTity of Tlie PCNETT to tWe 
hobbyist community, we made it possible 
to negotiate the independence of . the 
boundaries at those two levels. In 
addition, the default choice of that 
option is to align the boundaries so 
that coroutines are never required. 

The tension 
more highly c 
escape notice 
with its largel 
and more high 
transmission bl 
levels handle 
messages, so 
lower levels, 
message server 
are more likely 
expanded and so 
whose meanings 
the oeople main 

between symbolic data and 
oded data can scarcely 

in a system like PCNET, 
y symbolic message header 
ly coded stream and link 
ock headers. The lower 

blocks much smaller than 

header compactness is 

more important at the 

Programs implementing the 

level, on the other hand, 

to have thetr functions 

are more in need of data 
are easily discerned by 
taining them. 



BOX 1579, PALO ALTO CA 94302 

Sizarre though it may seem, PCKFT 
messages probably vill normally be sent 
from tfteir source lirectly to their 
destination or to a snail drop n*»ar the 
iestination. That is, they will 
orobably only rarely be sent through a 
lumber of internediate nodes on their 
rfay to their destination. (The reason 
is a simple economic one: the late night 
:>hon» rates are low enou7b and 
independent enough of distance that 
relaying doesn't seem worth the bother.) 
^wever, the PCNST specification is 
organize:? in such a way that either mode 
can be used, that any mixture of the 
nodes can he in use at any moment/ and 
that the choice can be made 
independently by the sender each time it 
attempts to send a message, in case the 
first few attempts fail. The crux of 
the matter is that the connection from 
the original source of a message to its 
ultimate destination can be a simplex 
(i.e., one-way) connection, aven though 
the connections at lower levels must be 
at least half duplex. 3n the other 
hand, there is nothing to orevent the 
original sender and the ultimate 
destination from having a full dual ex 
connection; in fact they can profit from 
having such a connection if they have 

The PCNET is 
network partic 
too: successful 
can be verifi 
destination al 
original source 
of the original 
of intermediate 
the transmissi 
their needin 
transmission b 
Since interais 
nominal respons 
they handle, t 
a streamlined 
considerable r 
spontaneous dis 

largely immune to nodes 

entering and leaving 

ipation in another way 

transmission of messages 

ed from the ultimate 

1 the way back to the 

(This is at the option 

sender.) Thus, a series 

nodes can participate in 

on of a message without 

g to accept any 

for ultimately successful 

etween the end points. 

diate nodes have only 

ibility for the traffic 

he PCNET should gain both 

administration and 

esistance to any possible 

appearance of nodes. 

s « l2Q£lUS.iOQ£ 

The ?ao st 
throughout th 
naximize its ac 
as possible a 
community. A 
that the desi 
that are inclui 
oe i nip lea; ant a 1 
is also structu 
sinple iniplen-? 
reliable. On t 
highly saeciali 
available or th 
structured in 
advanced facil 
deliver increa 
user. The desi 
the addition of 

which ker- 

i sic or t a p. t consideration 
e "CN£T ort.sign is to 
cessibi! ity to as bro*- 1 

segment of the hobbyist 
consequence of this i? 
gn his soecial f eaturp.s 
^d just so the design can 

simply and cheaply. It 
red in such a way that 
ntations will tend to b-» 
he other hani, when mor? 
zed cr costly hardware is 
e software can be mor-» 
than the rcinimuni 
the PCNET .iesicn is 
such a way as to put the 
ities to good use ani 
sed performance to the 
^n also leaves room for 
nev features, not all 0* 
r e s e e n b v the P c n 7 " 

5. ici^noaiaijaffisals 

Some of the opinions expressed in 
this paper are lay own. However, the 
vast majority of the ooinions, 




incorporated here w»re developed wholly 
or jointly by the members of the PCNET 
Committee. Hoping to find a reasonable 
middle ground between the folly of 
avaluatinq the work: of others and the 
arrogance of leaving their contribution 1 ; 
unacknowledged, 1 feel compelled to say 
that I found most memorable t s e 
contributions of Oave Caulkins, ?on 
Crane, Peter Deutsch, fr'arc Kaufman and 
Robert Maas. 

7 - Sifeli22£a2tiiial»&«lsiiac.2£ 

CI J. Oav*d Caulkins, "?esi-7n 
Considerations for a Hobbyist Coniouter 
Vetwork", Proceedings of the First Vest 
:oast Computer Paire, Palo Alto, 1977. 

C?l. Mike wilber, M A Network o* 
Community Information Exchanges: Issues 
and Problems", ?roc=edinas of the First 
West Coast Computer 'vaire, 3 alo Uto, 



BOX 1579, PALO ALTO CA 94302 

Communication Protocols for a Personal Computer Network 

Ron Crane 
2101 Calif. St. #326 Mountain View, Ca. 94040 

This paper summarizes the design requirements and 
architecture of a layered set of communication 
protocols for personal computers. Also included is 
the current state of development of this evolving 
set of protocols. A detailed description of the 
protocol will appear in the future. 

The Environment of Computers 

The introduction of low-cost minicomputers in the 
1960*s reduced the minimum size of an application 
for which automation was justified. The single 
chip microcomputer is making an equally 
significant step in the 1970's. Private individuals 
are purchasing small machines or terminals for 
their own use, in addition to businesses for 
applications such as accounting, order handling, 
and text editing. Small, medium, and large size 
computers are becoming widespread. 

More and more machines are being dedicated to 
uses which involve transactions of some kind 
(editing text, mail, or programs ultimately sent to 
other individuals, doing accounting for transactions 
between buyers and sellers, etc.). The machines are 
used to create, modify, or process information 
which is then printed on paper or recorded on 
magnetic media which is then carried to its 
destination where it is often entered into another 
machine. This represents a growing environment 
in which machines are communicating with other 
machines, but with the transport time from one 
mac hi ne tn another compr ising a significant part 
of the total elapsed time for processing. The need 
exists for low-cost, quick, and direct machine-to- 
machine communication between an increasingly 
widespread base of machines. 

The requirement of providing communication 
between any of a large number of machines implies 
some form of large and widespread communication 
network. To be a viable replacement or adjunct to 
physical transportation schemes presently used, the 
network must be available and relatively 
inexpensive. The only system currently available 
that is both widespread and has a very low 
minimum monthly charge is the dial telephone 

For machines to communicate using the telephone 
network, they must be able to communicate both 
with the telephone system and through the 
telephone system to the distant machine. Once two 
machines have been connected via the telephone 
system, they must communicate with each other. 
This communication must take place on several 
levels. Modems communicate using a standard set 
of frequencies, the serial interfaces in each of the 

computers must transmit and receive bits in the 
same order and at the same speed, and the software 
in each of the machines must agree upon the 
meaning of various bit sequences so that 
information is transferred instead of a meaningless 
sequence of bits. A common language or 
communication standard is essential for widespread 
machine communication. 

Personal Computer Network (PC Net) Protocols 

The PC Net protocols come from the PC Net 
committee which is a group of computer 
professionals who banded together as the result of 
the personal computing network session at the First 
West Coast Computer Faire in 1977. 

The PC Net protocols are layered into 5 functional 
levels. Two of the levels specify hardware while 
the remaining three specify software. Every 
machine must implement the basic or core 
protocol. Extra capabilities may also be 
implemented on various machines. The 

compatibility of two machines with respect to extra 
capabilities is ascertained on a per call basis via the 
core protocol, as is the decision to use one or more 
of the capabilities during the call. Thus, any two 
machines can always communicate with each other 
on a very simple level, even though some will find 
that they have few common interests after a brief 

The figure on the following page illustrates the 
~ hferarchicat structure of the proposed standards. 
Five levels comprise the standard (protocol). 

Level 1 

Level 1 specifies how the node (personal computer) 
interfaces to the telephone network. Both 
supervisory signals (dialing) and the modem 
signalling frequencies are specified. Bell 103 
compatible modem frequencies were selected 
because of wide availability and relatively low cost. 
When faster units become desirable, this level of 
the protocols can be changed without affecting the 
other levels. In the figure, RS232C and RS366 are 
mentioned as often-used interfaces to modems and 
auto-dialers. These are not part of the standard, 
however, since neither interface can be seen by 
another node through the telephone network. 

Level 2 

The data transmission format specified by this 
level is asynchronous start-stop codes at 300 bits 
per second, again readily available in existing 
equipment. The interface to and control of the 



BOX 1579, PALO ALTO CA 94302 







4 Auto 




Dial Telephone 





& Auto 





Bell 103 Modem Standard & 
Automatic Dialer 


Asynchronous Start-Stop Codes @ 300 BPS & Auto Dialer 

Link Transmission Blocks (sequenced and error free byte delivery) and 
Telephone line control 



Level 1 

Level 2 


Process to Process Stream Blocks (multiplexing of data link between user programs) 


Level 3 

Level 4 

User Program to User Program Standards (Node Control, Mail, File Transfers, etc. 


Level 5 

* RS232C and RS366 specify typica, modem and auto-dialer interfaces, respectively, but are not part of the PC Net 
standard since tne auto-dMer may be integral to the modem **** may ,n tun, oe part of tne computer interface. 

Personal Computing Network Protocol Hierarchy 



BOX 1579, PALO ALTO CA 94302 

Communication Protocols for a Personal Computer Network 

automatic dial and answer telephone interface is 
not specified except that the function must be 
performed. This includes the possibility of manual 
operation. Timeouts will be set to allow manual 

Level 3 

This level provides sequenced and error free 
delivery of bytes from one machine to another, as 
well as control of telephone connection setup and 
takedown. Blocks of bytes are transmitted on the 
link with a header and checksum in a fashion 
similar to synchronous protocols like ADCCP 
(Advanced Data Communication Control 
Procedure). A synchronous bit oriented protocol 
was not used because of the unavailability of 
hardware interfaces to implement it. Error control 
is implemented using a checksum and byte count 
instead of a CRC because the CRC is cumbersome 
in software and most asynchronous interfaces do 
not support it in hardware. Sequence numbers 
provide acknowledgement of correctly received 
blocks and in addition permit several blocks to be 
sent without acknowledgement, without getting out 
of order. 

Two methods of transmission on the line are 
specified. Radix-41 is the default start up mode 
and full 8-bit codes are an option. 

Radix-41 is a 2 byte to 3 character packing scheme 
proposed by Mike Wilber to avoid the problem of 
operating system intervention (interrupting on 
control characters and uppercase conversion of 
lower case characters) when PC Net protocols are 
implemented in BASIC or . other .. high level 
language. The result of using this packing scheme 
is a reduction of 33% in throughput. 

Transmission blocks are transmitted with only 
minor modification when using 8-bit codes. 
Special flag characters are used to separate 
transmission blocks on the link. A transparency 
rule is then used to permit transmission of this flag 
character if it occurs within the transmission block. 

Level 4 

This level is useful for nodes in which several user 
programs are running concurrently and are also all 
using the phone link. Level 4 performs 
multiplexing and flow control of data streams from 
each of the programs into the single phone link. 
Separate 8 bit fields specify source and destination 
process addresses for each process-to-process 
stream. Fields also exist for sequence and stream 
numbers. In small nodes this level can be merged 
with level 3. The fields specified for the level 4 
block format still remain, but can be ignored by 
the simple nodes. Small nodes can still 
communicate with large nodes capable of 

multiplexing, but only one process to process 
stream can exist at a time in this case. 

Level 5 

User programs exist at this level. They manage the 
communication tasks such as sending or forwarding 
mail or files, and interacting with people on whose 
behalf the communication is taking place. There is 
currently a mail and file transfer protocol written 
by Peter Deutsch and a global addressing scheme 
using latitude, longitude, and phone numbers 
authored by Doug Bourne. 


The PC Net protocols are being designed to be easy 
to implement in currently available systems, but 
with room for growth and modification if there is 
demand for it in the future. The first two levels 
are essentially complete at this time. The top three 
levels have been specified and are undergoing 
revisions relating to interfaces between these levels. 
Detailed descriptions of each level of the protocol 
will appear in the future. 



BOX 1579, PALO ALTO CA 94302 


(prepared from the online file PR0T0.PR3) 

by Robert Elton Maas (REM at SU-AI, MIT-MC) 

PO Box 6641, Stanford, CA 94305 

This document is one more attempt to 
explain the PCNet protocols to persons planning to 
write their own software. This document is 
written bottom-up so that node volunteers don't 
have to read it backwards to know what to 
implement in sequence. It also contains no 
completely worked-out examples (see [Maas WR2] 
for them, as well as a top-down approach), and no 
test data for debugging node software (see [Maas 
EXPERI]). It also doesn't fully specify timeouts and 
other obscure features of our protocols (see 
[Crane]; point of nomenclature - references of the 
form AAAA.BBB are to PCNet documents 
maintained as on-line files).). The two earlier 
primers, PROTO.PRIMER and PR0T0PR2 may or may 
not be useful to supplment this tutorial. 



This section will be mostly meaningless until 
later sections have been read, but due to our 
present indecision as to what certain things should 
be called there are some synonyms that ought to 
be pointed out before proceeding. 

Phone-line: modulated carrier, Bell 103 

Modem cable (optional): two-level bit-serial 
with start and stop bits, RS-232, not present if 
UART+MODEM on one board. 

I/O or memory-mapped hardware interface: 
8-bit bytes containing Radix-41 characters and 
framing, making up LTBs (Link Transmission Blocks). 

Link-level midpoint: binary translation of 
LTBs, called "TBs" (Transmission Blocks). 

Pure-binary stream: data portion of TBs, 
containing PPSBs (Process-to-Process Stream 
Blocks) packed end to end. 

PPSs (dynamically created and destroyed): 
data portion of PPSBs that comprise that particular 

Disk I/O: implementation dependent and 

(Note, the term "block" used by itself in this 
tutorial usually refers to a TB or LTB.) 


Hardware (not detailed below): 

For compatibility with existing 

software-service bureaus, Arpanet -host dial up 
equipment, manually operated terminals such as my 
Beehive used to remotely test software, and 
low-cost available personal computer serial-i/o 
interfaces — our network will initially be geared to 
the Bell- 103 standard rather than any of the 
brand-new bit-synchronous communications 
standards such as HDLC, and will run at 300 baud 
(110 optional) rather than 1200 baud. 

The interface to the main body of software 
will consist of a GETCHARACTER routine that 
returns -1 if no character has arrived at the UART 
(Universal Asynchronous Receiver/Transmitter) 
within a reasonable time and a number between 
and octal 377 (the ASCII, or rarely EBCDIC, value 
of the characer) otherwise, and a PUTCHARACTER 
routine that either waits until the character can be 
stuffed into the serializer or returns an error flag if 
it couldn't be stuffed immediately (wait -and- stuff 
would be used in halfduplex protocol, 
stuff-or-error would be used in full duplex 
protocol). An alternative is fully-buffered i/o that 
is driven by interrupts, but with similar interface 
characteristics. Timeouts are not a 

fulltime-eseential feature, rather are used to avoid 
telephone calls longer than necessary to either 
transmit a message or determine that the other 
node is sick, and to detect that a rare failure of 
halfduplex turnaround has occurred so that the 
deadlock can be resolved. It is expected that 
during 99% of connections not a single half-duplex 
deadlock will occur, thus with semi-manual 
operation all timeouts are optional. 

No further discussion of the hardware and 
its interface will be done in the rest of this 

An important segment of the PCNet protocol 
relates to Telephone Call Management(TCM). One 
of the PCNet design goals is graceful sharing of a 
single telephone line between voice and data use. 
TCM covers the activities which must take place 
from the time the PCNet node phone line goes 
off-hook until the time the Frequency Shift Key 
(FSK) handshake between the two communicating 
PCNet modems is complete. TCM also deals with 
call termination - the activities from the end of 
data transmission until the phone line goes on-hook. 



BOX 1579, PALO ALTO CA 94302 

PCNet TCM is intended for use in three different 
modes; 1) Attended (manual control). This mode 
assumes people present who will answer all calls, 
switching PCNet calls to the computer when they 
occur. 2) Attended (computer control). This mode 
assumes people are present for voice calls, but 
that the computer answers all calls and signals the 
people to accept voice calls. 3) Unattended. This 
mode assumes that people are either absent or 
asleep; the computer answers all calls, minimizing 
audible ring signals inside the cailed premises. 

TCM assumes thai a sophisticated modem is 
available; one capable of detecting telephone 
signals such as ring, busy, dial tone, etc. Simpler 
modems with time-outs may be used with the 
penalty of slower and less efficient operation. 

In present attended (computer control) and 
unattended TCM modes the computer goes 
off-hook for all incoming calls; if the call is voice 
and not data both people on the premises and the 
calling party must be given a ringing signal so that 
the voice connection may be completed. This is 
undesirable in that billing will start as soon as the 
phone line goes off-hook and before the voice call 
really begins. Also a special piece of equipment is 
required to generate the ringing signals. Some 
better way of differentiating between voice and 
data calls on the same phone line would be 

Link level (bottom half of communication software): 
Errors in transmission invariably occur when 
using modems over dialup lines. Thus a facility for 
detecting errors, requesting retransmission 

(implicitly PCNeL by a lack of affirmative 

response within a reasonable time), holding 
out-of-sequence blocks until the missing earlier 
blocks can be retransmitted, putting blocks into 
their correct sequence, and delivering verified and 
sequenced data up to the next level; has been 
included. In the default mode of operation, only 
one block can be sent at a time, eliminating the 
need for buffering and sequencing. 

When implementing our protocols as user 
programs on existing computer systems, limitations 
in the characterset available as input to a node, 
and sometimes even as output, usually make it 
impossible to transmit arbitrary 8-bit bytes across 
the line. For example, most systems ignore the 
parity bit on input, and set the parity on output 
regardless of what the programmer actually 
supplied as the octal 200 bit of outgoing data. 
Furthermore many systems supply linefeed to the 
input stream after carriage return that comes in, 
ignore null, do strange things like killing or holding 
output when control-0 or control-B is input, and 
interrupt the program completely when receiving 

control-C or control-Z. Some systems even 
convert all lower-case characters (octal 141 to 
172) into upper-case! To avoid almost any 
possible conflict when passing binary data, a subset 
of 41 (decimal) characters called Radix-41 has 
been selected to be transmitted across the line. 
Two 8-bit bytes of binary data are represented by 
three bytes of Radix-41. A more complete 
discussion of the alternatives, and reasons for our 
belief that Radix-41 is the optimal method for our 
purposes, are in [Maas RAD41]. 

Many systems echo back anything typed at 
their input lines, and on most systems this echoing 
cannot be turned off completely. Also, most 
programmers using BASIC on personal computers 
are unable to handle multiple concurrent processes, 
nor fully-buffered i/o, thus a fullduplex mode of 
operations where data is simultaneously travelling 
in both directions at the same time, is infeasible. 
We have thus chosen a method of simulating 
half-duplex mode of operation on any full-duplex 
or echoplex or true-halfduplex line, and it is the 
default mode of operation. To almost eliminate 
deadlocks caused by the turnaround character 
being lost due to line noise, 3-out-of-5 majority 
logic is used to determine whether the other node 
has or hasn't finished transmitting. To avoid being 
confused by seeing your own echo, a different 
turnaround character is used for the two nodes in a 
link ("[" vs. "]">• Also, each TB contains a bit 
telling who sent it, so that echoed TBs won't be 
accepted by the node that sent them in the first 
place (in the event that echo is delayed due to 
buffering, turnaround get momentarily confused, or 
echo occurs while in full-duplex mode). 

Node level (third quarter of communication 
software); ~ '"" 

The interface between the Link level and 
the Node level consists of a "pure binary stream" 
which is a sequence of 8-bit bytes with all 256 
possible values legal. Files can be transmitted 
using only the hardware and Link levels, however 
that leaves no method for starting and stopping a 
transmission. The Node level is a way to multiplex 
control information (start and stop of transmissions 
of files, as well as node-identification, network 
statistics reports, requests for fullduplex mode of 
operation, and advanced features we haven't even 
thought of yet) and data (one or more 
file-transfers, and maybe even additional services 
such as "TELNET" (link between a terminal and a 
remote interactive program)) along a single 
pure-binary stream. 

Virtual circuits are established between a 
process on one node and a process on the other 
node, for example between a process having a file 
it wants to transmit and a process that accepts 
files and saves them on its disk. We call such a 



BOX 1579, PALO ALTO CA 94302 

circuit, at this level, a Process-to-Process Stream. 
Normally here will be one PPS in each direction so 
that the two processes can talk in duplex Each 
PPS is broken into blocks, called PPSBs, so that it 
is easy to detect end of transfer without knowing 
the totai data count at the very si art (there is a 
bit in the header of each PPSB telling whether it is 
the last or not, and a count of zero data bytes is 
legal so that end-of-transmission can be sent at 
the very last possible moment if it is only then 
when the sending node realizes it really is done), 
and so that control messages and/or blocks of 
some other PPS can be multiplexed between 
PPSBs of any PPS in progress. 

Optional features include keeping track of 
buffer allocation at this level so that if one PPS 
gets blocked the link can still be used to transmit 
other PPSs without losing data of the blocked PPS 
(i.e. no data from blocked PPS is put into the 
pure-binary stream until the other node announces 
that it has more room in its buffer to store it). 

In the default mode of operation, only one 
transfer can be active at a time, no accounting for 
allocation of buffers is done, and any control 
messages or requests for additional transfers are 
ignored while a transfer is in progress. Also, the 
beginning of each PPSB is located at the beginning 
of a LTB and each PPSB is wholly contained in that 
LTB, so that an incoming block can be handled 
completely by a closed subroutine that preserves 
almost no state-information from one block to the 

The result of all this is that pure-binary 
duplex communications are established between 
File Transfer Process (FTP)/MAIL 

server-processes on the two nodes, without 
preventing upward compatibility into multiple 
concurrent services occurring between a pair of 
advanced nodes, and without preventing an 
advanced node from talking to a very-simplest 
node. In the simplest implementations it is 
expected that all levels in the receive pipeline will 
be compressed into a single piece of code that 
checks incoming data for validity, discards anything 
containing transmission errors, discards any PPSB it 
isn't able to accept, ignores any control mesage it 
doesn't understand, and depending on a boolean 
variable either awaits a request for opening a file 
transfer or delivers data from the active transfer 
to some output file. The transmit pipeline would 
need a queueing mechanism so that replies from 
the FTP/MAIL server-process could be sent as 
soon as turnaround occurs, and TBs containing no 
data could be sent if no data is awaiting 

Server-process level (top quarter of 
communication software): 

A server-process is a subroutine or other 

chunk of program which actually does something 
useful, as contrasted with all the routines at lower 
levels that do the grungy stuff necessary to make 
server-processes possible. Thus the 

server-process level is the highest level of the 
protocol. At present we define two 

server-processes, "control" and "FTP/MAIL". A 
server-process is specific to one particular type of 
activity, whereas all the other levels of software 
are general -purpose for linking nodes and 
server-processes together. A "listener" is that 
part of a server-process which sits waiting for the 
other node to give it a request to do something, as 
contrasted with performing an action after the 
request has been received, or spontaneously 
initiating actions or requests, which are done by 
non-listener parts of a server-process. 

The control-listener handles any PPSB 
addressed to process 0, checking to see if it knows 
the control opcode, and if so then acting on it 
somehow. In advanced nodes, an explicit 
negative-acknowledgment will be issued for 
anything it understands but rejects, and for 
anything it doesn't understand that is in affirmative 
mode. A simple node can ignore anything it doesn't 

The control process also has the duty of 
generating control messages to be sent back. 
Semi-mandatory is identification of the node. 
Optional are requests to go fullduplex or binary 
<non-radix41) or unsynchronized (PPSBs and LTBs 
overlapped for increased thruput) or allocated 
(explicit allocation granted before data can be sent 
on a non-control message, except for initial 
requests for transfer which are "borrowed" from 
the later allocations). 

The FTP/MAIL listener handles any new PPS 
that is addressed to process 1. These usually 
consist of requests for starting a file transfer. If 
the request is acceptable at this time, a duplex 
connection (one PPS in the other direction, in 
addition to the already-started PPS that requested 
the transfer) is established between a piece of 
code on one node and one on the other node. 
When the end of the original PPS occurs, 
end-of-transfer is signalled, and an 
acknowledgment and optional file-checksum is sent 
on the reverse PPS just before it too is closed 

Local forwarding/mailing software outside 

Before a message or file can be transmitted 
to another node, it must be created somehow, and 
then queued for the server-process to handle. 
The queued version must contain a PCNet header 
telling where its ultimate destination is, and may 
also have in the header a decision as to where the 
next hop will be (i.e. to which node it will be 
forwarded next). The PCNet communication 



BOX 1579. PALO ALTO CA 94302 

software must then somehow get started, either by 
answering the phone, or by the system or another 
program or a human explicitly starting it up. 

After a message has been received and the 
phone hung up (or possibly after the message has 
been received but while the phone link is being 
used for additional messages or other services, if 
the computer system is powerful enough to do so 
many things at the same time without slowing down 
the phone link), it must be decided whether the 
message is for local delivery or must be forwarded 
to yet another node. If forward, go to preceeding 
paragraph, else some delivery method must be 
chosen, such as appending to the start or end of a 
person's message file, writing as a new file with 
header stripped off and an anouncement of its 
arrival appended to the mail file, listing in hardcopy 
for manual delivery, or summoning an operator to 
telephone the addressee and recite it verbally. 

Since all this occurs outside the telephone 
link, it can be handled at leisure by whatever 
combination of manual and automatic means the 
owners/operators of the node may decide upon. 
However, once the server-process accepts an 
FTP/MAIL request, it is mandatory that either an 
explicit negative-acknowlegment be given, or the 
message or file be correctly delivered to its 
intended recipient. Any node which accepts 
messages and then loses them without comment, 
will get nasty black marks in the 
master-network-directory, and other nodes will be 
warned not to trust the offending node, which will 
probably result in ostracism until the offending 
node is fixed. To avoid this, if an operator 
discovers that his system has crashed and 
obliterated from the disk some message entrusted 
to it, he (she) should-lmmediately telephone the 
original sender to inform him (her) of the accident 
so that he (she) can re-submit the message to the 
network. (This of course implies that each node 
have some form of backup, such as a hardcopy log 
of message headers.) 


Radix-41 transmissions: 

The mapping between the 41 characters of 
the Radix-41 characterset and the values to 40, 
is nicely diagrammed in [Wilber, page 41], but can 
be summarized briefly. Open parenthesis, "(", maps 
to the value 0. Numerical-digit characters "0" thru 
"9" map to the values 1 thru 10 (note the offset, 
character "5" maps to value 6 for example). 
Uppercase alphabetic characters "A" thru "Z" map 
to values 1 1 thru 36. M *" maps to 37, "+" to 38, 
"- H to 39, and T to 40. 

To be sure the reader understands the 
Radix-41 representation, the most commonly 

misunderstood part of our protocols, I will now 
present algorithms for converting to and from 

Three characters of Radix-41, corresponding 
to the three numerical values they map to, are 
combined to make one 16-bit number. The first 
numeric value is multiplied by 41*41, the second 
by 41, and the third by 1, then these three 
products are added. (More efficient is to compute 
V3 + 41*(V2 + 4UV1), which only uses two 
multiplications and two additions.) Going the other 
way, two divisions must be done somehow (the 
most efficient way is either a table lookup on 4 
4-bit-byte8 extracted from the 16-bit quantity, 
adding the four value-triples looked up in the four 
tables, or for each division a multiplication by an 
approximation to 1/41 followed by a slight 

To convert from Radix-41 characters to 
binary, first map Radix-41 to numeric values, then 
do the two multiplications to get a 16-bit number, 
then break that number into two 8-bit bytes. The 
mapping can be done efficiently by range checks 
for numbers and upper-case-letters (in either case 
a simple subtraction of a constant will finish the 
mapping) followed by a check for the other five 
characters if the range checks fail. 

To convert from binary to Radix-41, first 
combine the two bytes into a 16-bit quantity 
(unless you are using the 4-table lookup method), 
then do the two divisions by 41, using the 
remainders and the final quotient as the numeric 
values in reverse order (i.e. the final quotient is 
the first value, the second remainder is the second 
value, the first remainder is the third value), and 
finally map to Radix-41 by a simple index into a 
table of 41 characters. 

All this conversion between binary and 
radix-41 actually occurs as subroutines in about 
the second-next section in this tutorial, however 
the algorithms were discussed here so the reader 
understands what Radix-41 is and is not before 
getting embroiled in halfduplex turnaround and LTB 

(Note, an alternative to Radix-41, 
transparant 8-bit mode, is described in [Crane]. It 
will not be explained in this tutorial. The 
discussions below of HDX turnaround, LTBs, and 
dropping of the odd byte, apply only to the 
Radix-41 mode of transmission. Other sections 
apply equally to both modes of transmission.) 

Half-DupleX (HDX) turnaround: 

To avoid deadlocks when one or two 
characters at the end of a transmission is (are) 
garbaged, a 3-of-5 majority rule is used. Five 
turnaround characters are transmitted, and as soon 
as at least three of them have been received at 
the other end, that node begins transmitting 



BOX 1579, PALO ALTO CA 94302 

without waiting for idle line or the rest of the 
turnaround characters. REM made an arbitrary 
choice of open and close square bracket (in ASCII, 
octal codes 133 and 135, gotten by shift-K and 
shift-M on most terminals even if not shown on the 
key), which has been tentatively accepted by 
everyone in the protocol committee, and so has 
been written into our protocols. (If we discover a 
high rate of errors, and find two other characters 
not already used by our protocols but having much 
lower error-rate, we are willing to change this 

To avoid overlapping an LTB with the tail of 
the half-duplex turnaround from the preceeding 
transmission in the reverse direction, or even in 
the same direction if the node starting to listen will 
see its own echo multiplexed with the beginning of 
the other node's transmission, each transmission 
begins with a series of five (or more) at signs. 
(Atsigns were chosen, after considerable 
experimentation, for their remarkable property of 
fixing UART character-misframing, thus assuring 
correct UART sync for the LTBs that follow, even if 
due to hardware or interface problems the 
beginning of a transmission has the UART in some 
funny state such as transmitting half a character 
then being reset and immediately starting another 
character, or interleaving incoming data and echo of 
one's own outgoing data at the bit level (like my 
Beehive does when in halfduplex mode!!!)). 

Finally, I will answer one of the most 
frequently-asked questions, which I always have to 
look up myself. Which node transmits "[" and 
which node transmits "]" The answer is that node 
(the node that originated the call) transmits "]" and 
node 1 (the node that answered the phone when it 
rang, which implies an answering modem) transmits 
"[". Thus after- node 1 answers the phone and 
starts up the PCNet communication-program, it 
transmits [[[[[ to indicate it is listening for the first 
actual transmission, then node transmits @@@@@ 
followed by one or more Link Transmision Blocks 
(LTBs) followed by ]]]]]. Then node 1 transmits 
@@<a@@ followed by its blocks followed by [[[[[. 
This alternation continues until one or the other 
node hangs up after either agreeing with the other 
node that they are done, or getting disgusted with 
the other node and giving up. 

Link Transmission Blocks (LTBs); 

In addition to the mapping from binary into 
Radix-41, each LTB is prefixed by an atsign, and 
followed by another atsign. Thus a total of six (or 
more) consecutive atsigns occur at the beginning of 
each transmission, adjacent LTBs are separated by 
two atsigns, and one atsign occurs immediately 
before half-duplex turnaround brackets. Atsigns 
serve two purposes, correcting UART mis-frame, 
and delimiting LTBs at the software level (in fact at 

THIS level right here). Of the two atsigns between 
LTBs, the first one fixes the UART so the second 
will get through to software, and the second one 
(or the first if both get through) tells the software 
to finish any preceeding LTB that it was parsing 
and to start another as soon as a valid Radix-41 
character (not atsign, not brackets) occurs. 

An LTB as actually transmitted, therefore, 
consists of one atsign, 3*N Radix-41 characters 
(representing 2*N 8-bit bytes that are one TB), 
and one more atsign. 

After a LTB (delimited by atsigns) is parsed, 
it is mapped into binary using the subroutine 
detailed earlier, if its length is not a multiple of 3, 
the mapping fails and it is rejected immediatey. 
Also if it is exactly or 3 characters long (the 
former check is an easy way to determine the 
nothing between two consecutive atsigns that 
occur between LTBs, namely you pretend it is an 
LTB then reject it because it is too short to really 
be one). 

If timing is critical, it may be necessary to 
buffer up a complete transmission without checking 
validity of LTBs, or even without parsing them at 
all. All checking can be postponed until after 
half-duplex turnaround. On the PET, using BASIC, 
even that trick loses, and it is necessary to make 
the main loop buffer up incoming characters 
without even checking for turnaround-brackets. 
When the main loop detects that no character has 
arrived, it lets a second loop, a co-routine, steal a 
moment of CPU time to perform one step in a loop 
that searches the buffer for turnaround- brackets. 
(Two routines are co-routines, as contrasted with 
one main routine and one sub-routine, if each 
returns to where the other left off, rather than one 
of them (the sub-routine) always being restarted.) 
Thus during the actual transmission the HDX-search 
co-routine lags far behind, then has time to catch 
up while the line is idle after HDX turnaround, at 
which time it finally sees the brackets and signals 

Odd/Even checksum: 

This check, and the next three, can be done 
in any of the twenty-four possible sequences. The 
order shown here is random, but based on 
heuristics (actually prejudices). In actual fact the 
most efficient sequence is probably (1) 
drop-odd-byte (2) compare length (3) check 
orig/ans (4) compare checksums. 

All the odd-numbered bytes are added up, 
modulo 256, and compared with zero. All the 
even-numbered bytes are added up and checked in 
the same way. (This can be done in a single loop 
by swapping the two running sums and always 
adding to the same variable.) If either result Is 
nonzero, the block is rejected. 

When transmitting a block, it's a little more 



BOX 1579, PALO ALTO CA 94302 

tricky because the two checksums obtained must 
be complemented before appending to the TB, and 
the appending must be "reversed" if there is an 
odd number of bytes so that the receiving node 
will get zero for each alternating checksum. See 
below for more on the extra zero byte added after 
the checksum when the TB has an odd number of 

Originate/Answer flag: 

The octal 200 bit in the first byte of the TB 
is the originate/answer flag. It equals the node 
number of the sender of the block, which is the 
complement of the receiving node number. 
(O=originate l=answer, referring to original 
establishing of connection as well as to modem 
frequencies. Nodes not using Bell 103 protocol, for 
example radio links, must agree who is node and 
who is node 1 before blocks can be exchanged.) 

When receiving a block; if this bit isn't the 
complement of one's own node number, the block is 

Drop odd bvte: 

If the TB-length, which is the complete 
second 8-bit byte of the TB, is odd, there is an 
extra zero byte (the last byte from Radix-41 to 
binary conversion, which is the next byte after 
TB-length has been exhausted) at the end of the 
TB which must be ignored. The easiest way to 
handle this is to decrease-by-one the local 
variable that tells the actual size of of the TB 
decoded from Radix-41 (before this step it will 
always be even, and if this step is performed it 
will then be odd like the TB-length in the 
second-byte already was). 

tt is reccommended, if you can afford the 
code, to first check this byte to be sure it really is 
zero, and to report an error if this check fails after 
the other three checks have succeeded. 


The length as specified in the second byte 
should now equal the received-TB-size minus two 
(because the checksum bytes and the extra zero 
byte aren't counted, and the extra zero byte has 
already been deducted). If this step is performed 
before removing the extra zero byte, then the 
NUMBER is compared with the received-TB-size, 
instead. if the checksum has already been 
computed, and discarded from then 
ZERO BYTE IF TB-LENGTH IS ODD, then you don't 
even have to offset by two, the two counts will 
exactly agree. In any case, when receiving a block, 
if the lengths disagree the block is rejected (thus 
two blocks concatenated, whose checksums will 
always combine to yield zero, or a block which has 

been truncated but whose received part is all zero 
thus adding to zero, will be rejected. Even a 
steady stream of zero values, represented by the 
open parenthesis character, will be rejected, 
because the TB-length in the header of the TB will 
be zero whereas it should be at least two. Even if 
it it is two due to noise on the line, it is unlikely 
that a checksum of exactly 376 to offset it could 
be created by the same burst of noise without 
messing up the other interleaved parity.). 

Sequencing of blocks modulo 8: 

The first two bytes of each TB are the 
header. We've discussed everything in it except 
the two sequence fields. One (called SEQ) is the 
sequence number (of the TB it is in) modulo 8 and 
the other (called REVNAK in our current software 
~ this term as well as SEQ, RCVNAK, XMTNAK and 
XMTGEN may be changed in later documentation 
and software) is a sequence number for 
not-yet-received blocks traveling in the reverse 
direction. The latter is of interest to the transmit 
pipeline when we receive a TB containing it, and is 
taken from a globally-available parameter in the 
receive pipeline when transmitting a TB. As it sits 
inside (first byte, mask 007) the TB, I call it 
REVNAK which means "REVerse Negative 
AcKnowledgement". It comes from the RCVNAK 
variable of the transmitting node, and is stored in 
the XMTNAK variable of the node that receives it. 
Thus it is affiliated with the REVerse process to 
the one that is actually passing it across the phone 

The SEQ field refers to the TB it is in, 
identifying its sequence number modulo 8 (i.e. a 
3-bit number). When constructing a TB for 
transmission, a local variable XMTGEN is used to 
generate this field, then XMTGEN is increased by 1 
(modulo 8) to be ready for the next TB to be 
constructed. The XMTGEN variable in each node 
starts at zero, thus the blocks sent by each node 
are numbered 0,1,2,3,4,5,6,7,0,1,2,3,... This block 
number, a 3-bit field in the TB header (octal mask 
070 in the first byte), is then used by the 
receiving node to decide whether it has already 
gotten it, is expecting it now, or isn't yet ready for 
it. The first and third cases are generally 
indistinguishable. In either case the block is 
ignored except that since the checksum etc. all are 
ok it is known to be A BLOCK from the other node 
hence the REVNAK field can safely be copied to 
the XMTNAK variable the same as if it was a block 
that was accepted. (Normally the storing of 
REVNAK into XMTNAK is done just before the block 
sequence number is checked, but of course after 
all parity+length+extrazero+originate checking is 

If the received block is numbered exactly 
the same as RCVNAK (oldest still-not-received 



BOX 1579. PALO ALTO CA 94302 

block, initialized to zero) then the header and 
checksum are stripped away and what's left is 
passed up to the pure-binary stream (if the TB 
length was exactly two, then there are exactly 
zero bytes of data, so nothing is passed up) and 
the RCVNAK is increased by on9 and reduced 
modulo 8. These two steps, passing up data and 
updating RCVNAK, must be properly synchronized 
(if this software has multiple processes active at 
the same time) so that it is impossible to get two 
copies of the same block and accept them both. 

If a node is capable of buffering more than 
one block at a time, then it is possible to .get an 
acceptable block that isn't the very next one but 
which can be saved at the receiving node until the 
intervening one(s) get (re)transmitted to fill the 
gap. Then when the one matching RCVNAK finally 
arrives, after updating RCVNAK another check is 
made to see if the one matching the new value of 
RCVNAK has already arrived, in which case it too 
can be unbuffered and passed up to the 
pure-binary stream, RCVNAK updated again, and 
the test for already-arrived-RCVNAK-equal block 
repeated until it finally fails. (At most four blocks 
can be receivable at one time, thus at most three 
blocks already buffered can be emitted when the 
one before them finally appears. This limit of 
4-out-of-S is absolute, there is a counterexample 
(or scenerio) that demonstrates that with a window 
of 5 or more out of 8, a block can be mistaken for 
one 8 earlier or later and thus completely destroy 
function of the link. This fact won't be apparent 
until we've finished discussing the function of the 
REVNAK field and the XMTNAK variable.) 

The function of the RCVNAK variable, as 
observed from the outside (in particular from the 
transmit pipeline which may run asynchronously at 
the same time as the receive pipeline when in 
full-duplex mode) is now describable. At any 
moment it equals the number of the oldest 
not-yet-received-and-completely-processed block. 

Each time the transmit pipeline sends a 
block, at the last possible moment before 
transmission (just before converting to Radix-41), 
it copies the RCVNAK variable into the REVNAK 
field of the header, and recomputes the checksums. 
(Note this also occurs on retransmission of a block, 
thus it is always current in realtime at a moment 
just before the start of the LTB output.) The 
effect is that the REVNAK field is an implicit 
acknowledgement of all blocks (in the other 
direction) preceding the one numbered REVNAK, is 
an explicit negative acknowledgement about the 
one numbered exactly REVNAK, and is a "no 
comment" on all later blocks that might have been 
received and buffered or might not have yet "been 

We may define the time that an 
acknowledgement really happens to be the instant 

when the node receiving the block, after checking 
parity etc., stores the REVNAK field into the 
XMTNAK variable. Thus at any moment, the 
XMTNAK variable in a node equals the sequence 
number (modulo 8) of the oldest outgoing block 
that hasn't yet been acknowledged by the other 
node. When XMTNAK equals XMTGEN, it means 
that all outgoing blocks constructed earlier in this 
session have been sent and acknowledged. 
Otherwise it means the blocks numbered from 
XMTNAK up to but not including XMTGEN are either 
somewhere on the round trip out (as blocks) and 
back (as implied acknowledgements), or have been 
lost somewhere due to line noise or lost characters 
due to too-slow software or other problems. A 
heuristic algorithm in the transmit pipeline uses 
XMTNAK and XMTGEN as well as any other 
information available, to decide whether to 
retransmit a block that still hasn't been 
acknowledged, transmit a new one while waiting a 
little longer for the wayward block to get 
acknowledged, or perform half-duplex turnaround. 
In the simplest implementation, which is the default, 
a window of l-out-of-8 combined with half-duplex 
mode makes the choice obvious, namely upon 
getting turnaround the program checks XMTNAK 
against XMTGEN. If XMTNAK+1=XMTGEN (mod 8) it 
retransmits XMTNAK block. If XMTNAK=XMTGEN it 
constructs a new block from queued data (if none 
there's some hair, consult REM and look at a listing 
of the program PCNSR3.SAI for details). The other 
six possibilities are impossible, thus indicate 
programing bugs or faulty hardware inside one 
node or the other. 

Pure-binary stream: 

The data portion of TBs arrives as a stream 
of bytes (or as successive arrayfuls of bytes, one 
array from each TB). It may be fed directly to the 
PPSB parser, or buffered first. In any case there 
should be an explicit control point through which ail 
incoming pure-binary data passes, and another 
similar control point for the transmit pipeline, so 
that software below this level can be made almost 
totally independent of software above this level, 
and so that during debugging a trace can be 
installed to see what is getting thru these two 
points (rev and xmt pure-binary) in the software. 
Why break the software at these two points? 
First, because it is the only place in the entire 
software, other than the UART interface, where a 
conceptually-clean place gets 100% of the data 
flowing through the software. Second, it is a 
natural place to splice different protocols together. 
During initial testing of the bottom half of the 
protocol it is common to put the software in 
loopback mode by feeding ail incoming data right 
back out. Slightly later the lower half can be used 
almost-stand-alone for file-transfer using a simple 



BOX 1579. PALO ALTO CA 94302 

top-half kludge (halter-top?) that is initiated 
manually at each end or semi-automatically by a 
restriction on the characterset and a convention for 
detecting end of file. (The program PCNSR2.SAI is 
an example of this.) Another possibility is to 
replace the bottom half by some commercial data 
network that guarantees 1007. perfect 
transmissions of 8-bit bytes, or by the bottom half 
of the DIALNET protocol, or by one of the new 
bit -synchronous communications protocols. 

PPS Blocks: 

At the node level, data is explicitly broken 
up into PPSBs, each of which is one segment of a 
PPS (Process-to-Process Stream). If necessary 
these can be multiplexed (complete PPSBs from 
one PPS located between those from another PPS, 
but a PPSB isn't broken internally, thus a PPS can 
be broken only at PPSB boundaries). The very 
first byte of data in the very first LTB, during any 
given session, is also the first byte of header of 
the first PPSB in that session (this applies 
separately to the two pipelines, one in each 
direction, transmit and receive). Once the pipelines 
have started, the PPSB-LENGTH field in each PPSB 
determines where it ends, and the next PPSB 
starts immediately after (the next byte of the 
pure-binary stream). Since the data in the 
pure-binary stream is 100% accurate, no errors can 
happen in parsing (except by failure of the 
equipment, which is assumed to not happen), so no 
other framing is required. In the default mode, with 
LTBs and PPSBs synchronized, there is additional 
redundancy that can be used to detect such 
"never-happen" failures. 

The first PPSB in any PPS has two fields not 
in later -PPSBs, namely a SOURCE-PROCESS and 
DESTINATION-PROCESS number, each 8 bits. The 
DESTINATION-PROCESS field is used to direct the 
PPS to the correct piece of code to start 
processing the PPS. After the PPS has started, 
that code passes control to whoever will really be 
reading the PPS, which may be a 
specially-generated entry point, or a throw-away 
sink if the PPS isn't acceptable for receipt. The 
SOURCE-PROCESS field is used in case the 
DESTINATION-PROCESS wants to open a PPS in the 
reverse direction, it uses the SOURCE-PROCESS of 
the original PPS to fill the DESTINATION-PROCESS 
field of the reply PPS. 

Every PPSB has the following fields: 
BLOCK-LENGTH (used to parse PPSBs in 
not-delimited end-to-end format of the pure-binary 
stream), PPS-NUMBER (used to identify which PPS 
it is part of), BLK-NUMBER (sequence number 
within that PPS, starting at for the first-PPSB 
which has the two extra fields described above), 
and LAST-BLOCK-BIT (1 if this is the last block of 
the PPS, so that the PPS it is in will be formally 

closed at the end of this PPSB, otherwise the PPS 
will be kept open waiting for more PPSBs). 

Normally the block number is almost 
redundant. If it is zero for a PPSB that isn't part 
of a known PPS, it is assumed to be block (rather 
than block 8 etc.), and the DESTINATION-PROCESS 
field is checked for legal values (0 and 1 currently, 
except when it is a reply to an FTP), otherwise it 
is a non-first block and can be ignored. Note that 
at most 8 PPSBs can be sent in a new PPS before 
getting positive confirmation that the other node 
has actually accepted the PPS, otherwise block 8 
might be confused with a new block attempting 
to open a new PPS. 

In advanced nodes with multiple link-levels 
over separate dialup lines to increase effective 
bandwidth, the PPSBs for one PPS can be 
distributed to different link-levels, and the block 
numbers used for reassembling them in correct 
sequence. But this is not likely to happen for quite 
a while! 

Note that if the process originating 
(SOURCing) a PPS realizes after it has already sent 
the last byte in a PPSB with LAST-BLOCK-BIT 
zero, that the PPS should now be closed, it can 
simply send a PPSB with BLOCK-LENGTH equal to 
exactly 3 (i.e. 3 bytes of header and bytes of 
data) and with LAST-BLOCK-BIT set. Thus it is 
never too late to close an open PPS. There are 
also methods to abort a PPS rather than closing it, 
usually to signal some error condition. This 
capability is included in the control messages listed 
in [Maas WR2]. 

Process-Process Stream: 

After deciding where to send the data part 
^-rr^SBrtn^-strtpping off the 3-word or 5-word 
header (5 for the first PPSB in a PPS, 3 for later 
PPSBs in the same PPS), the remainder of the 
PPSB (up to 255-3=252 or 255-5=250 bytes, 
minimum 5-5=0 or 3-3=0 bytes) is passed up to 
whereever it was supposed to go, to be 
interpreted further by whatever program is located 
there. Two such programs, the Control -listener, 
and the FTP/MAIL server, are defined presently. 

ProcesssO — meaning of link-control messages: 

The Control -listener gets any PPS that is 
addressed to destination process 0. Each PPS 
must consist of exactly one PPSB, with the 
LAST-BLOCK bit turned on (1). Any multi-PPSB 
addressed to process is considered a violation of 
protocol. Each PPSB to the Control-listener 
consists of exactly one control message. The first 
byte (of PPSB-DATA) is the control opcode, and 
any remaining bytes are arguments. A list of 
currently designed control messages are in [Maas 
WR2]. Almost all them can be ignored safely by 
unfancy nodes. 



BOX 1579. PALO ALTO CA 94302 

Process*! — protocol for transferring a file 
(message or other); 

1 are fed initially to the FTP/MAIL listener. After 
a PPS has been opened, a siightiy different entry 
address will probably be set up so that the rest of 
the PPS can be processed without constantly 
rechecking for same/different PPS number. Usually 
a dispatch table for currently-open PPSs will exist 
in the PPSB parser, and simply changing an entry in 
it will switch a stream to a different piece of code 
without affecting anythng else. Any new PPS 
addressed to process 1 will go to the original 
FTP/MAIL listener, which will then note that an 
FTP/MAIL is already in progress and either ignore 
or signal abort unless the computer can handle 
multiple simultaneous file transfers (a rare situation 
even on large computers). 

The general protocol for transferring a file is 
as follows: The node wishing to send the file first 
sends a IFTP (Initialize File-Transfer-Process) 
request to process 1 of the node that will (maybe) 
receive it. That node then checks to see if the file 
is small enough to fit in available storage, and if not 
replies with a NO-I-WON'T IFTP message. If, 
however, the IFTP is acceptable, it replies with an 
l-WILL IFTP message. Then on the same PPS as 
the original request, the sending node transmits an 
followed by actual data of the file all the way to 
the end of the PPS which is then closed. Finally a 
reply of either l-WILL TAKE-THIS-FILE containing a 
checksum of the data received (zero if no 
checksum was computed) if it was successful, or 
l-WONT TAKE-THIS-FILE with some error code if 
not. The receiving process should actually safely 
close the file it is writing into storage, before 
sending the l-WILL ... <checksum> confirmation of 
completely successful transmission. After the 
transmitting node gets the confirmation, and 
verifies the checksum is correct or zero, it may 
safely delete the copy it was reading from. 

Local interface between FTP/MAIL-server and 
mail-forwarding queue; 

This is mostly up to the operators of the 

Local mail-forwarding program: 

General guidelines for forwarding of 
messages are in [Deutsch], [Bourn], and [Maas] 
(respectively, MAIL/FTP protocol, 

worldwide-addressing, and final -deli very). How 
forwarding is actually accomplished is up to the 
individual operator(s). 

Local mail-creating/editing/receiving program(s): 

When creating a message or initializing an 

FTP, the correct header should be put at the start 
of the file. The forwarder and the protocol 
program can then handle it as a chunk of data, 
examining the header when necessary to see what 
to do with it next. When delivering the message 
to the addressee, it is optional how much of the 
header to keep, how much to edit to make it 
prettier, and how much to purge. All this is up to 
the operator(s) of the system. 


We hope that this tutorial has helped the 
reader understand most of internal workings of the 
program that at one end of a phone connection 
automatically maps files down through the layers of 
protocol and transmits them out the telephone line, 
and at the other end receives the LTBs and maps 
them up through the layers to construct a copy of 
the file. Formal specifications of these 

communication protocols are given in [Crane]. 
Specifications of how a message-forwarding 
network is built upon these protocols are given in 
[Wilber], [Bourn] and [Maas WR2]. These and 
other documentation and tutorials from the PCNet 
Committee are available at a nominal charge to 
cover the cost of reproduction and mailing. 
(Contact the author, or any of Dave Caulkins, Mike 
Wilber, Ron Crane or Peter Deutsch.) 

At the time of final-editing of this tutorial 
(1978 January 16), protocols are working on 
several PDP-10 computers, and this software has 
been partially transferred to the PET (using BASIC) 
and to the Altair (using assembly language). It is 
hoped that we can have several micro-processor 
nodes fully-working and doing useful 
electronic-mail service by the time this paper is 
presented at the fair. 


[Bourn] "A Proposal for Addressing Stations of the 
Personal Computing Network", by Doug Bourn, an 
internal working paper of the PCNet Protocol 
Committee, available by photocopy means only. 

[Crane] "PC Net Communiation Protocols", by Ron 
Crane et al, a working paper of the PCNet Protocol 
Committee, in press (online file PR0T0.PB1). 

[Deutsch] "Mail Transfer and Forwarding", by Peter 
Deutsch, a working paper of the PCNet Protocol 
Committee, in press (online file SERPRO.TTY). 

[Maas] "Final Delivery of Messages", by Robert 
Maas, an internal working paper of the PCNet 
Protocol Committee. 



BOX 1579, PALO ALTO CA 94302 

[Maas EXPERI] "Experiments Sub-Committee of 
PCNet Committee - Status Report", by Robert 
Maas, continually-updated file (online file EXPER.I). 

[Maas FLO] "Flowcharts Showing Data and Control 
Interfaces between Modules in PCNet 
Protocol/Software", by Robert Maas, in press 
(online file PROTO.FLO). 

[Maas RAD41] "Explanation of Radix-41 in PCNet", 
by Robert Maas, in press (online file RAD41.WRU). 

[Maas WR2] "The Design of the Personal-Computer 
Network (PCNET)", by Robert Maas, in press 
(online file PROTO.WR2). 

[Wilber] "A Design for a Network of Community 
Information Exchanges", Mike Wilber, presented at 
the first West Coast Computer Fair e. 


Disneyland on your doorstep? 

Jim Dunion, The American Museum of Energy., P.O. Box 117, Oak Ridge, TN 


The American Museum of Energy in 
Oak Ridge, Tennessee has recently begun 
a program to introduce and develop micro 
computer technology in a museum environ- 

The museum has some data processing 
requirements, which, although not very 
elaborate, form certain pre-requisites 
for additional computer activities. 

Micro-computers are being util- 
ized in museum exhibits in two ways. 
Some are complete, stand-alone , terminal 
oriented exhibits. These exhibits 
allow certain energy topics to be intro- 
duced to the public, as well as provide 
a means for receiving direct feedback 
as to public opinion. Longer range 
plans call for building micro-computers 
directly into certain interactive 
exhibits . 

Beyond the direct utilization of 
nicros in exhibits, the museum is an 
active resource for computer technology. 
This is accomplished in several ways: 
oy offering beginning classes in micro- 
computer technology and programming, by 
sponsoring computer activity events at 
the museum, and by actively soliciting 
and promoting community participation 
In museum activities and developmental 
pro j ects . 

Moving Micros into the Museum 

The American Museum of Energy in 
Dak Ridge, Tennessee (AME) , has recent- 
Ly initiated a program to introduce and 
Jevelop micro-computer technology in a 
nuseum environment. The need for such 
a program ties in very closely with the 
current energy situation in the United 
States. The AME operates under a 
contract from the Department of Energy, 
and has as one of it's main functions 
to increase public understanding of 
science and technology, with particu- 
lar emphasis upon energy issues. Since 
the energy program that President Carter 
announced contains a major shift in 
emphasis away from developmental energy 
programs towards fossil fuel usage and 

conservation methods, the need to 
educate the public as to the real 
energy situation and what they can do 
about it is more urgent than ever. 
It is the feeling of the museum staff 
that microcomputers can play a major 
role in increasing the effectiveness 
of this educational process. 

Studies performed at this museum 
and others indicate that a great deal 
of care must be taken in planning 
exhibits if they are to successfully 
transfer meaningful information to a 
visitor. The average time spent at an 
unattended exhibit varies between 30 
seconds and several minutes. Thus 
exhibits cannot rely too heavily on 
written text, but must incorporate 
interesting graphics and audio-visual 
presentations. The current tendency 
at science and technology centers is 
to design interactive exhibits with 
which visitors can participate. Most 
exhibits currently manufactured employ 
hard-wired circuitry and electro- 
mechanical devices, making them expen- 
sive to design, implement and change. 
Again, the feeling at AME is that 
micro-computers can make a significant 
impact on exhibit and display technol- 
ogy. So, the AME decided to take the 
bull by the horns and instigate a 
program to develop and implement this 

Who's Involved 

The primary 
behind this move 
the chairman of 
Oak Ridge Associ 
is well acquaint 
technology, havi 
of the Lawrence 
Berkeley. Durin 
was responsible 
Hall's public ac 
program which h 
tremely successf 

However, at 
Science, as well 
and technology c 
almost exclusive 

motivating force 
is Robert F. Content, 
the Museum Division of 
ated Universities. He 
ed with computer 
ng served as asst, dir. 
Hall of Science at 
g his tenure there, he 
for developing the 
cess to computers 
as proved to be ex- 
the Lawrence Hall of 
as at most science 
enters, there is an 
reliance on mini- 



BOX 1579, PALO ALTO CA 94302 

computers. It is now time to put micros 
in the museum. 

The first step in this project was 
to acquire some computer equipment. The 
museum already had two Wang 2200's, and 
a PDP-8, but these are utilized for in- 
ternal data processing and in a couple 
of existing exhibits. To get things 
going, four SOL ' s from Processor Tech- 
nology and two Silent 700 terminals from 
Texas Instruments were purchased. 

Next, a specialist in microcomputers 
and personal computing, Jim Dunion, was 
hired to head up this developing program. 
He brought with him a Compucolor 8001 
with a floppy disk. At that point, the 
museum faced a somewhat unusual situation, 
equipment rich-people and software poor. 
With that in mind, the watchword of our 
program became - cooperative development. 
In a nutshell, we hope to solicit help 
from the local community in Oak Ridge and 
Knoxville in developing microcomputer 
technology in the museum. We provide 
the equipment , the educational training, 
and the projects. In turn, we hope to 
receive time and software from interested 
individuals . 

Major Areas of the Program 

We have broken our program down into 
four main areas. 

1. Social 

2. Educational 

3. Developmental 

4. Communal 

The social aspect of our program is 
intended to generate interest in the 
museum's computer activities. A" little 
showmanship, if you will. We have 
started a series of computer activities 
nights at the museum. These sessions 
are aimed towards familiarizing the gen- 
eral public with the capabilities and 
benefits of personal computing. Each 
session is keyed around a central theme, 
such as the recreational uses of computers 
or computers and art. A typical night 
will consist of a movie about some aspect 
of computers, a technical talk (in terms 
that laymen can understand) , several 
systems running demonstration programs, 
and refreshments. 

Our educational program will proceed 
in several discrete steps. Initially we 
are holding a two to three hour seminar 
called, An Introduction to Personal 
Computing. After presenting this talk 
several times, we will initiate two 
short courses, 

1. Understanding Personal 

2 . Beginning Programming in 

After these classes have been 
conducted a few times, we will then 
begin offering more specialized topics 
such as: 

1. Game Playing with 

2. Advanced Programming 

3. Special Projects 

4. Computers for Kids 
One of our goals for this part 

of our program is to interest local 
programmers and personal computing 
enthusiasts in teaching these and 
other courses. Hopefully also, some 
of the students that we train can 
then turn around and teach future 
classes . 

The developmental portion of 
this program promises to be one of the 
more interesting aspects of our 
approach. We have identified four 
major research areas of interest for 
developing micro-computer technology 
in the museum. We have begun acquiring 
additional hardware to configure 
systems in different work stations 
for these projects. As local students, 
hobbyists or just interested indiv- 
uals begin working with us, we will 
coordinate special assignments in 
these research areas to provide some 
training and experience for the 
workers, and of course, exhibits for 
the museum. The research areas are: 

1. Verbal Information 
Systems ' 

2. Color Graphic Display 


4. Automated Exhibit 

Verbal Information Systems 

The primary goal of this project 
is to develop an energy information 
system that anyone could walk up to 
and ask certain questions about energy 
technology or energy policy issues. 
The catch is, we want the input to be 
spoken language, and the output to be 
synthesized speech, clearly not a 
trivial task. This project will involv 
voice recognition, speech synthesis, 
natural language understanding, data 
base management, and artificial intell 
igence. In addition to a SOL computer 
we will add a voice recognition unit 
and a speech synthesized for the 
initial system configuration. 



BOX 1579, PALO ALTO CA 94302 

Co lor Graphic Display Technology 

One of the systems we have avail- 
able is a Compucolor 8001, which is a 
color graphic device. We already have 
a number of interesting games and demon- 
strations for this system, but the 
feeling of everyone who sees this sys- 
tem is that we are just scratching the 
surface. We need to develop programs 
that make the creation of displays and 
games easier. Such programs might 
include a complete graphics package 
(vector graphics are already provided) , 
a rubber-band drawing system, an ani- 
mation package, etc. 

We already have some specific 
exhibits in mind to use this system. 
Particularly in light of the fact that 
Intelligent Systems Corporation, who 
manufactures the Compucolor, is prepar- 
ing to announce the Compucolor II, a 
scaled down version rumored to sell for 
under $1000.00.; We plan to place an 
orientation device in the lobby of the 
museum that will present certain infor- 
mation about the museum. One thing that 
will be displayed will be a floor plan 
of the museum. Visitors will be allowed 
to specify certain rooms of the museum 
on the display (probably using a light 
pen) , and then receive more information 
about the exhibits in that room. We 
feel that if visitors have an overview 
of the museum before they start their 
tour then the visit will be more mean- 
ingful . 

Another use for this system will 
be to provide interactive simulations 
of such things as global energy systems, 
household energy conservation, nuclear 
reactions, etc. 

Finally, we are designing several 
games oriented around energy. These 
are right now called The Energy Game 
and Embargo. 

The hardware for this project 
area is already complete. 

Electronic Bulletin Boards 

The Electronic Bulletin Board will 
provide a message posting system, text 
editing and word processing, newsletter 
preparation and information about on- 
going projects (both energy and compu- 
ter related) . 

We are currently looking at text 
editing and word processing software 
for this project. In addition, we 
plan to install a telephone interface 
so that the system may be accessed 
remotely. Currently, our plans call for 

using one system to act as the phone 
access and also to control a hard 
disk mass memory device. This in 
turn will interface to the other 
systems for both remote access and 
downline loading of programs. 

Automated Exhibit Technology 

The purpose of thi 
be to provide expertise 
electro-mechanical devi 
this is the most specul 
projects. Our first goa 
develop a small, inexp 
to work with. We will 
the basic design presen 
Heiserman. We also wan 
a visual input system, 
we will be very active 
United States Robotics 

s project will 
in controlling 
ces. Right now 
ative of our 
1 will be to 
ensive robot 
probably use 
ted by David 
t to install 
Naturally , 
in the 
Society . 

As is probably apparent from 
the description of these research 
projects, we hope to make them 
exciting and interesting enough so 
we can attract a lot of local talent 
to work with us. 

The Museum as a Community Resource 

As part of our program to attract 
co-workers, the museum is actively 
promoting the idea that we are a 
community resource. We are doing this 
in several ways. First, we are 
sponsoring and supporting local 
computer hobbyist groups. There is a 
club in Knoxville that has been active 
for some time now, and a club is just 
forming in Oak Ridge. The facilities 
of the museum are available (to some 
degree) to these groups. The actual 
mechanisms of this availability will 
have to be worked out as the program 
proceeds . 

We have also initiated a lecture 
program that is geared towards local 
civic groups. The main topic is 
Computers and Energy (with a huge 
plug thrown in about the museums 
computer activities, naturally). 

Finally, we are trying to acquire 
as many publications in the personal 
computing field as possible, so that 
they may be available for local 
hobbyists. We are also gathering as 
much software as we can, software in 
the public domain that is, and will 
transfer this software to interested 
parties. We are particularly inter- 
ested in software relating to energy 
or simulations related to energy. 



BOX 1579. PALO ALTO CA 94302 


The museum's program is still in 
the developing stage. One thing we want 
to emphasize is that we very much want 
and need as much help as possible in this 
endeavor. Advice, software, hardware,... 
are actively solicited. We hope to 
expand this into a national program, and 
towards this end we will be publishing 
a newsletter tying together the personal 
computer movement and the Association 
of Science and Technology Centers (a 
group of 80-100 such institutions) . This 
newsletter will be available to hobbyist 
clubs, schools, etc. 

The impact of micro-computers on 
our everyday life continues to grow. 
We think that their introduction into 
a museum environment will be very 
significant indeed. Perhaps in a few 
years museum exhibits will resemble 
Disneyland rides. With your help, and 
our efforts, we'll put micros in the 



David and Annie Fox 

Co-Directors of Marin Computer Center 

70 Skyview Terrace Room 301 

San Rafael, CA 94903 


Marin Computer Center is a project of Ulenar, a non-profit, educational 
corporation, whose main goal is to bring the wonders of advanced technology (computers 
and the like) within the reach of all people. 

We have set up 10 microcomputers in what was formerly the library of Oakview School 
in San Rafael, California. In a spacous, well-lit room, with beamed ceiling, orange 
carpeting a many plants, we've created the kind of comfortable environment that has 
never before been associated with computers. 

We will describe how MCC came to be, what it is, and where we plan to take it. 

How 1 1 A I I Began 

Marin Computer Center was seen as a vision at first. We came upon the idea - or it 
found us - quite unexpectedly in mid-August of 1976. How strange it seems now, and yet 
very natural all at the same time. 

To say that computers and the world they represented was far from the world that we 
inhabited then would be a gross understatement. At that time in our lives, and for 
several years prior to that time, we were "spiritualists" - lovers of the occult, 
psychic realm - followers of numerous "personal growth" excursions - always seeking. 
We considered ourselves very much the "humanists" - with our respective careers of 
teaching and counseling. 

We felt that not enough people were coming in contact with new ideas about 
themselves, not enough people were growing in their personal lives. The question was, 
how to introduce the vast majority of Americans to themselves. We took a look around 
and noticed the beginning boom of video games. What if we developed a video game in 
which people could learn more about themselves and their relationships with others in 
the process of playing? Of course, the stated purpose of the game wouldn't be personal 
growth, that would just be a side effect of playing it. 

From this idea we jumped to a fantasy of a huge complex similar to Disneyland. The 
main difference would be in the participation level of the visitors. Disneyland is fun 
but it is essentially a place where they "do it to you". You watch animated dolls while 
riding on a boat or go for a submarine ride and watch sea serpents looming at you. No 
one is given an opportunity to interact with the environment, to play with the 
environment in a way where some new and unique learning experience would result. We 
envisioned a technology pi ay I and where all this could happen. To actually "person" the 
deck of the USS Enterprise with other visitors and make contact with other worlds. To 
warp your own intergalatic vessel around the universe while looking through a three 
dimensional viewscreen and experiencing the force of acceleration. To feel 
weightlessness in a zero gravity room. The movies "Westworld" and "Futureworld" are the 
closest we've seen to this idea. Of course, the conflicts of man versus machine in 
those films represent the fears we wanted to help people overcome in order to make the 
most of technology. 

With our long range goals set, we had to find something which we could accomplish 
with today's technology. The concept of the Marin Computer Center was born. We 
embarked — whole-heartedly without a backward glance. It seemed as if we had been 
running full steam in one direction - then one day screeched to a halt for no externally 
apparent reason - and zoomed off at twice the velocity down a new road! 

It may seem strange that two people with no technical background would be audacious 
enough to enter the hallowed grounds of "computer land", but somehow our naivete has 
served to make the whole thing unique and appealing in the eyes of others. 

We created Marin Computer Center because we felt that there needed to be some 
educational facility that would bridge the gap between peoples' fears and their natural 


curiosity about computers. It seemed evident to us that the rapid growth of the 
personal computing industry would result in a "computer in every home" by the early 
1980's. Judging that as an inevitability and evaluating the prevailing attitude about 
computers, it seemed obvious that people needed a painless way to ease themselves into 
the Computer Age. 

Many people feel that computers are cold, dehumanizing instruments of 
totalitarianism. The image of Big Brother and the "Computerized Society" seem to go 
hand in hand. At least that has conventionally been the fictionalized view. We would 
be the first to admit that in the recent past computers have been used in ways that have 
resulted in general feelings of powerlessness and compartmental ization. However, it is 
important to distinguish between computers (the species) and how they've been used. In 
other words, it is short-sighted to condemn a device simply because of the misuse and 
abuse it has suffered at the hands of people with something less than the "common good" 
in mind. 

Alarmists and political paranoids argue that computers are potentially dangerous in 
that they can be used to store incredible amounts of very personal data and then recall 
that information at an astonishing rate. They become uneasy at the thought of the 
"Master Computer" controlled by the CIA. 

The Computer is a powerful tool. And it, like many powerful tools throughout 
history, has been used and misused by people who seek power for purposes of both good 
and ev i I . 

When the printing press was first invented, the church began to fear its use for 
the purpose of widespread propaganda against Church Doctrine. They launched their own 
campaign against the machine, condemning it as a tool of the Devil. One would have to 
admit that there have been some pretty libelous, degrading and socially unredeeming 
things that have been presented to millions of people in the form of the printed word. 
However, one would not be hard pressed to think of just a few of the beautiful, 
inspiring, and beneficial things we have experienced through our exposure to words in 

So which is it? Tool of the Devil or Invention of Enlightenment? Actually the 
printing press is neither. The prnting press is just a machine that prints words on 
paper. The discussion is arbitrary and meaningless. The same is true of the debate 
about the potential joys and evils of a computerized society. 

Computers are here to stay. And the general public needs to start taking 
responsibility for its own personal participation in the world of computers. Because 
they are such "all purpose" machines, it is up to us to decide which of their various 
purposes are ones that we want to support. 

Marin Computer Center's main goal is to "introduce people of a I I ages to computers 
and. -the advanced technology which they represent In order that anyone wight begin 
participating in the process of computer assistance for society". 

When we started we felt certain that our objectives were valid and would serve a 
valuable function in this society. But lofty goals and innovative plans are meaningless 
if they cannot be manifested in the physical universe. And in order for our dream to 
take a real form we needed money. 

Our quest for capital led us to dozens of private foundations. We spent six months 
peddling our grant proposals with no success to speak of. 

For long periods of time our goal seemed extremely distant and as likely as a 
winning sweepstakes ticket. In the face of such overwhelming odds and dispair, were we 
discouraged? We sure were! Weeks went by and nothing happened - no forward movement; 
our plan was stagnating and so were we. Many times it seemed as if we continued with 
our phone calls and letters just to spite all the people who thought we were crazy to 
persist with an idea that couldn't get off the ground. And I'm sure we must have been. 
Crazy enough to continue persisting even though the Foundations weren't exactly beating 
a path to our door, we knew it didn't mean that money couldn't be obtained through 
another source. 

So we did what most people do when they need money - we hit the banks. And lo and 
behold, with the help of a friend (with more financial credibility than we had) our loan 
application was approved! 

That was in July of 1977 - a full eleven months after the whole idea was hatched! 
In the two months that followed, we rented 5,000 square feet in a beautiful school 
building, ordered and received nine Sol-20's and one Equinox, obtained some programs, 
invited 300 people to an Open House, placed three ads in local newspapers - and held our 
breath . 


Visiting the Center 

On September 10th we opened our doors - at long last Marin Computer Center had 
crossd over into the physical universe. Over 700 people showed up for our Open House 
celebration and during the past six months they have steadily continued to come. Little 
children with their parents, neighborhood kids stopping in after school, handicapped 
children and adults, older people - all with their interest in computers to guide them. 

MCC was created to give people an experience of computers and the advanced 
technology which they represent. In the first few months that we've had our doors open, 
we have in fact been providing that kind of experience in addition to many other kinds 
of experiences that we had not anticipated. 

For example, on Saturdays MCC provides a place for families to come together in an 
attractive and calm environment for a unique "learning experience". We see them come 
in, wide-eyed and slightly apprehensive. They have heard about this place from friends 
of theirs (who had "a terrific time") so they thought they'd see for themselves. They 
don't have any idea what to expect and frankly, they've got their guard up. We greet 
them and make them feel welcome. We acknowledge the uncertainty they are exuding and 
they begin to feel that they don't have to pretend that they're feeling at ease when 
they're not - their anxiety is understood and then they begin to relax. 

We tell the newcomers about our set-up, in terms that they can relate to. We talk 
about why we've created this center and that we're glad that they've come to explore. 
After talking for a while, we suggest a computer game that might interest them, load the 
machine and let them settle in for the fun of confronting a new learning experience. 

. Adults and children relate to new learning situations in totally different ways. 
We have learned much from observing people with computers. Children seem to be very 
much attracted to the CRT terminal - because of their familiarity with TV and home video 
games, children between the ages of 7 and 10 feel very much at home with our 
microcomputers. Their attitudes towards the computers are open, eager and an almost 
matter-of-fact acceptance of the things that the technology of today has managed to 
accomplish. Older children, while equally open, seem to be more apreciative of the 
wonder of it all. They have reached a point in their own cognitive development to be 
able to imagine in abstract terms what a computer is and how it manages to do what it 
does. (There is a greater preponderance of 14 year old boys who frequent the center 
than any other age group.) So although children of different ages may be experiencing 
the computers differently, they all are unanimous in their enjoyment of and fearless 
approach to the machines. 

Adults, on the other hand, are less likely to welcome the challenge of this 
particular "unknown" with open minds. Adults come to the center with the whole gamut of 
preconceived attitudes, ideas and beliefs about computers. Their experience may have 
been in the form of a mistaken IRS refund, a cancelled magazine subscription that kept 
on coming or other annoyances that have been blamed on a "computer foul up". With these 
kinds of things in mind, many adults come to the computer center ready for a fight, it 
seems. They are sour-faced individuals who wish that the animal "computerectus" would 
go on the endangered species list and not survive. Then there are women in the 35-50 
age group who feel intimidated by the "superior" intelligence of computers. They are 
embarassed that the computer will make them look foolish by knowing more than they do. 
And finally there are the older adults (in the 50-70 age range) who are bewildered by it 
all. They feel that the world is just moving too quickly and that they are being left 

After a direct experience with computers, one's fears are seen as groundless. Then 
the individual creates the opportunity for him/herself to really explore the computer as 
a new personal medium of creative expression. 

One of the ways both adults and kids can do this is by taking one of our classes in 
computer programming. The class is really an introduction to microcomputers and the 
computer language BASIC. The course covers a brief history of computers - through 
vacuum tubes to transistors to integrated circuits to large scale integration; 
discussion of how a computer works and then right into learning the language and 
creating your own programs. 

The class is for absolute Beginners - no prior knowledge is assumed or expected. 
Since we personally tiptoed into the field without the usual prerequisites we fully 
understand and empathize with the fear and general uncertainty people bring with them 
into our classes. Because of this empathy we are particularly good at creating a safe 


learning environment for them to explore these "intelligent" machines. 

Graduates of our courses have gotten right into the process of using computers in 
their lives for more fun, profit, and efficiency. Some examples are: the man who 
created a program to calculate the milk production of his goats, the teacher who used 
the course to create specialized curriculum for his junior high school deaf students, 
the woman who was in charge of the reservations department in a large airlines and 
wanted to have more knowledge of computers to increase her feelings of effectiveness in 
her job and the 14 year old boy who has created computer programs for the games of 
Yahtzee and Battleship. 

One's success at survival has always been based on the ability to adapt; a 
willingness to change. With the world's increasing rate of change we've all got a 
chal lenge just to keep up with it. And more important than keeping up with it is to be 
a part of that process of change. We at Marin Computer Center are giving people a 
wonderful opportunity to participate in that area of change in today's world known as 
"Computers". By directly interacting with computers, people begin experiencing new 
feelings of freedom and confidence, replacing their former fear and confusing 

All of these people have an experience at the Computer Center which enables them to 
step outside of their preconditioned feelings of hostility, fear and confusion and enter 
a new world. A world that is not the de-humanizing robot world that they first imagined 
- but a world of people and learning and change instead. It's an exciting new world, 
and there is a place in it for everyone. The child in all of us is fascinated by 
computers - the "New Age Toy, Tool and Servant" of Humankind. 



BOX 1579, PALO ALTO CA 94302 


-r^: 1 

Ludwig Braun, Professor, Department of Technology and Society, 
SUNY at Stony Brock, Stony Brook, F.Y. II?9^, yi&-?.k6-%UlS 

Educational technology has gone from 
the log ( for Plato and his student 
to sit on) in i+OO BC through the 
book in lk^ when the Gutenberg Bible 
became the first mass-produced book, 
to the personal computer whose genesis 
is reckoned to be either January 1975 
when the Altair was announced, or 
April 1977 when the PET was announced 
(depending upon you definition and 
your loyalties). 

Until now, students and teachers 
were involved in information trans - 
mission ; while, now, they can begin 
to think in terms of information 
processing — with the enormously- 
increased intellect-enhancement this 

Because of the low cost and portabil- 
ity of computers like the PET and the 
TRS-80, educational computing sud- 
denly has become much more attractive 
than ever before. These computers 
are totally self-contained and need 
only an a-c outlet to operate. This 
means that learners can use computers 
anywhere without worrying about tele- 
phone locations, etc. The teacher 
can plan to bring the computer into 
the classroom or the office. Because 
personal computers are portable, 
society can think about putting them 
in neighborhood or school libraries 
so the kids ( or adults) can sign 
them out. The $600 price is low 
enough that parents can think ser- 
iously about giving their children 
computers for Christmas or as birth- 
day presents. 

The graphic capabilites of the PET 
and the TRS-80, although limited 
compared to very expensive graphics 
facilities, are very impressive at 
the price. Art teachers can give 
their students a new flexible medium 
to explore. One important advantage 
of the computer as an art medium is 
that the child can convert art in his 
mind into visual images without the 
manual dexterity required by clay, 
paints, crayons and other media. 


The capability to speak and to re- 
cognize spoken words (a. la Speech- 
lab and Oomputalker) provides s Items - 
tive input and output modes to permit 
children or handicapped people to use 
computers; and provides teachers of 
languages with computer support in 
a new and interesting way. 

One of the most exciting capabilities 
of personal computer?; (to me at least) 
is the ability to connect to the real 
world through analog-to-digital (A/D) 
and d igital -to -analog (D/A) converters. 
This capability permits us to develop 
simulations which provide much more 
realistic learning experiences than 
is possible by more traditional com- 
puter methods. With a/d and D/A con- 
verters, the digital computer becomes 
an analog computer with all the advan- 
tages of the digital computer and none 
of the disadvantages of real analog 

The most exciting possibility to im- 
prove learning environments is the 
combination of the personal computer 
and the almost-available video-disc 
system. Such systems have been called 
"intelligent video-disc systems" by 
Professor Bork of the University of 
California at Irvine. In Bork's con- 
cept, the computer and the video-disc 
player interact with each other and 
with the user. In such systems, the 
computer controls the video disc player, 
causing it to play a motion sequence, 
a set of still frames, an audio sequence 
or causing a computer program stored on 
the video disc to be loaded into the 
memory of the computer for subsequent 
execution. There is a feeling of ex- 
citement about the intelligent video- 
disc system, and the potential for 
dramatic impact on learning. Such 
systems are some time off (perhaps 3-5 
years), but will be worth the wait. 

The rate of development of personal 
computers and related peripherals is 
breathtaking. It is impossible to 
guess what new announcements will be 
made six months from now ( or even 
at the Second Computer Faire), but 
the consumer and especially the learner 
gains with every one. 
177 BOX 1 579. PALO ALTO CA 94302 


Arthur Luehrman 

Associate Director 

Lawrence Hall of Science 

University of California 

Berkeley CA 94720 



Science Museums have a unique opportunity to educate a broad public about the use of computers. Less than 10% 
of today's high school graduates have laid hands on a computer keyboard, and even fewer of their parents have done so. 
Yet the needs, in terms of jobs and personal development, for computer skills is growing rapidly. With 50 million visitors 
annually and ties to local schools, science museums can play a critical role, and a few have. For many years, the Lawrence 
Hall of Science has had a vigorous computer education program based on its 100-port time-shared computer,with terminals 
in about 50 Bay Area schools, on the exhibit floor and in the classrooms at the Hall. Inexpensive personal computers are 
expected to increase our potential impact tenfold or more in a few years. Among other projects to be described during the 
panel are these: (1) Putting a dozen computers in a van and driving to schools and clubs to conduct workshops, (2) Plug- 
ging into exhibits computers with programs that ask questions about the exhibit or suggest activities, (3) Offering classes in 
which students will be lent computers to take home and work with, and (4) selling computer and application programs in 
the museum store. 



BOX 1579, PALO ALTO CA 94302 

PANEL: Personal Computers and Learning Environments 
TOPIC: Computers for Elementary School Children 

Bob Albrecht 
P.O. Box 31 
Menlo Park, CA 94025 


They are finally here! The $600 plug-em-in-and-use-em What would I change to "improve" the PET? 

home/school computers are here! Not one, but two 

$600 computers are available! Wnat would ' chan § e t0 "improve" the TRS 80? 

The Commodore PET What do ' like about tne TRS 80? Dislike? 

The Tandy Radio Shack TRS 80 Wny aren > t there any easy-to-learn computers? 

Both are complete computers - the $600 price includes What wou)d you !ike t0 have in a nom e/school 

computer, memory, alpha numeric keyboard, video computer for kids? 

(TV) display and cassette recorder. 

Bye, bye paper Tape! 

Hello magnetic tape cassette! 

So, back to the classroom ... or the family room. For 
the next year or two or three I will be visiting elementary 
schools, helping students and teachers learn how to use, 
program and enjoy the PET and the TRS 80. You will 
frequently find me in the school learning center, 
resource center, or library where the computers will be 
an open resource, available to all students and teachers. 
Fifth and sixth grade students will learn to be 
"resource people" or "teacher aids". They will be 
available in the center to help others learn how to use 
the computer. 

Right now, there are very few programs available for 
our use, but people are working on that. There are 
no instructional materials, so I am developing "teach 
yourself" style materials to help students, parents 
and teachers learn to read and understand BASIC. 
Instead of the usual "math" approach, I am using a 
verbal and graphics approach. 

I will attempt to answer questions about teaching 
elementary school children how to use, program and 
enjoy the Commodore PET and the Tandy Radio 
Shack TRS 80 computers — at school, or at home. 
Questions mgiht include, but not be limited to 
the following. 

How do I teach my 4th grade child or student 
how to program in BASIC. 

How does an elementary school get started using 

How do we overcome teacher inertia? 

Where do I get education software? Instructional 

What do I like about the PET? Dislike? 



Liza Loop 
LO*OP Center, Inc. 
P.O.Box 9^5 
Cotati, CA 9^928 

Many teachers ask, " Why is it im- 
portant for school children to become 
aware of computers? " My answer is 

o The computer is the most signi- 
ficant technological innovation since 
the printing press. You may prefer to 
compare it to electricity. Either way, 
the computer is too important to be ig- 
nored by established education. 

o Although not all school child- 
ren will become computer scientists, 
they will all be consumers. They will 
have telephones, receive bills, and 
shop at supermarkets. Therefore, they 
will all be consumers of computer ser- 
vices. If nothing else, they need to 
understand that computer service systems 
are designed by people and can be chang- 
ed by people . 

o Most school children will vote 
at some time in their lives. In this 
capacity they will have to make deci- 
sions concerning acceptable and abusive 
uses of computer technology, including 
uses of computers by the government it- 

o Finally, most people will work 
during some period in their lives. 
Studies perdict that over 80$ of the 
jobs available in 1985 will require some 
involvement with computers or their in- 
put and output. Since schools are sup- 
posed to address the problem of prepar- 
ing youth to emerge into the world of 
adult responsibility, it is hard to jus- 
tify not teaching about computers. 

Most students can develop some 
sense of what a computer is and what it 
is not. Fourth graders will avidly play 
computer games and many will choose to 
write their own programs rather than ac- 
cept canned material . Even much younger 
children rapidly develop an ability to 
manipulate a terminal provided the out-» 
put is on their level. 

What is Computer Awareness? It is 
the study of the computer system itself 
and the place of computers in today's 
world. It is significantly different 
from computer assisted instruction (CAI] 
which uses the computer as a delivery 
system for a variety of non-computer 
subjects. Computer Awareness includes 
units on operating the terminal, using 
games and other application programs 
( possibly some CAI ) , computer applica- 
tions found in the locality of the 
school, computer related careers and vo- 
cations, introductory programming in 
high level language ( perhaps Basic ) . 
On the secondary level it may go on to 
cover some electronics and Boolean 
logic. Computer Awareness is not Com- 
puter Science, Programming, Data Proces- 
sing, Electronics, or Computer Maintain- 

Computer Awareness also differs 
from Computer Literacy. It omits the 
history of computers and deemphasizes 
skill in programming. It places more 
stress on experience with computer 
applications and less on vocabulary 
which is often meaningless to younger 
children and tends to be obsolete very 

Computer Awareness students do 
write at least one program. However, it 
may be as simple as a picture program 
using only PRINT statements. The exer- 
cize is intended to develop a sense of 
the relationship between the user and 
the programmer rather than a means to 
promote skill in programming. 

L0*0P Center, Inc. has tried a num- 
ber of different approaches to the sub- 
ject of computers in education. It ran 
a demonstration classroom and storefront 
drop-in computer center for two years. 
In September, 1977 » the storefront 
portion of L0*0P was closed in order to 
allow concentration on teaching the 
Computer Awareness Curriculum on site at 
various Sonoma County schools and to 
write a companion text book for the 



BOX 1579, PALO ALTO CA 94302 

In talking with school staff it be- 
came appearent that many teachers wanted 
a prepackaged curricular module which 
they could import into their classrooms 
with a minimum of effort. Others ex- 
pressed an independent interest in the 
computer field and shared in the belief 
that computers will become an integral 
part of organized education in the near 
future. L0*0P began to focus on the de- 
velopment of a Computer Awareness Pack- 
age and on presenting this material in 
courses for teachers at Sonoma State 

One constantly annoying obstacle 
to both LO*OP's storefront operation and 
to the development of an easily trans- 
portable Computer Awareness Curriculum 
was the lack of reliable, low-cost hard- 
ware. The Sonoma County Computer Club 
provided an on-going evaluation labora- 
tory for microcomputers as one member 
after another bought and struggled with 
kit computers which never quite lived up 
to manufacturers* claims of power and 
ease of operation. LO*OP borrowed and 
field tested many of them in classrooms 
only to spend most of one period search- 
ing for a disconnected wire or another 
watching paper tape load. 

The conclusion is that the Computer 
Awareness Package must include a fully 
debugged and assembled machine which 
speaks a high level language and comes 
up with not more than four or five in- 
structions from the keyboard. Currently 
under consideration are Radio Shack's 
TRS80 and the Commodore Pet. Both these 
machines speak reasonable Basic as soon 
as you turn the power on and cost about 
as much as a classroom movie projector. 

Within the next year, materials 
will be available so that every teacher 
may introduce Computer Awareness into 
his or her classroom. But, as we are 
all well aware, this project is only one 
tiny contribution in the field of com- 
puters in education. 



BOX 1579, PALO ALTO CA 94302 

Implications of personal computing 
for college learning activities 

^essqrch Scientist 

i'iie purpose of this article is to 
provide background inform ation for a 
session on "personal computers and 
1 earning environments 1 ' at the raire, a^-i 
..--erhaps more important, to stimulate 
continuing discussion of tnis topic 
among those who see the Proceedings, 
ihnt is said in the session builds on 
./hat is written here, but do not expect 
it to be the same. And six weeks after 
the raire we should expect to have new 
information, new ideas, and nev/ 
opportunities for expanded use of 
personal computing in college learning 
activities. Still this background 
information should be useful. 

The domain of personal computing I 
consider more broadly than just small, 
single-user machines. I visn to include 

some. Bxperiiinces with timesharing 

systems in oruer not to overlook useful 
ideas about communities of learners and 
occasions for professional communication 
within such communi ties. If we limit 
discussion to small (inexpensive, 
portable, individually owned) machines 
.ve are not taking advantage of decades 
■yf experience with personal computing 
using costly, fixed systems having the 
jower (processor speed, memory size, 
instruction set, and complexity) that 
Afill be characteristic of the small 
machine in a few years. I am convinced 
that the best personal computing is done 
today on single-user machines, but I 
haven't csiven up entirely on timesharing 
as a means to providing truely personal 
services. In any case, I want to have 
communication networks backing up 
single-user systems in education. 

The substance of this oackground 
statement is presented in five sections. 
The first is intended to provide the 
educator some indication of why the 
personal computing revolution is so 

important to computer use (and 
information nandling) in higher 
education. l he second section should 
inform the computer specialist or 
enthusiast aoout kinds of uses in higher 
education. ihe thiro. section offers a ( 
list of wnat I see to be needed to nelp 
along some noeoed changes in higher 
education. i'he last two sections 
provide a orief statement about the 
future an- a list of references and 
but jested readings on computers in 
nioner education. 

oiJMrlCAMCif i)r PcifSO/UL OOMPUTI .13 
i-\)K CdLLL-Jc LiiAtfNlUG 

Many colleges and universities nave 
acquired effective and economic 
time-sharing systems for use by faculty 
and students in teaching and learning 
activities. These can be expanded to 
accomodate tne increasing amounts of 
use. Also, they can oe extended to 
handle some new kinds of uses sucn as 
graphics ana information retrieval 
systems. However, some kinds of 
computing will be accomplished more 
economically with small and inexpensive 
computers for use by one individual or 
project at a tine. At the University o f 
Michigan the Center for Research on 
Learning ana Teaching is exploring 
various roles for these microprocessors 
gnu personal computers within college 
learning activities. 

Aspects of the revolution 

i'ne impact of personal computers on 
education in general and instructional 
activity in ^articular will oe 
considerable. Three aspects of tne 
p e r 3 o i \a I c oi.ip u t e r r s v o 1 u t i on ma y he 1 p 



BOX 1579, PALO ALTO CA 94302 

establish tiii.') case: numbers, -iccess, 

.illy Ci.'i ( K,l Oj. • 

personal computers wi LI be used in 
mucn larger nuiiiuers t:wr: instructional 
computers ivjve been or will be 
otherwise. i'he uajor computer-based 
instruction facility so far nas been tne 
PLATO Computer-based Instruction dystem' 
at tne University of Illinois in w.-iich 
the number of simultaneous users are 
counted in tne hundreds. Jhanvies in the 
system will expand ttie number of users 
to thousands, i'he :;icst common 
timesharing systems in use in colleges 
toaa/ are ouilt by Kewle t t-Packnru and 
jigital equipment. Although these are 
small systems (four to 60 users each), 
the total numoer of simultaneous users 
served 0/ the huncreos of systems 
throughout trie country is counted in the 
thousands. with continuing sales to 
colleges this number •■/ill expand to the 
tens of thousands. At least' I 00 ti;;ies 
lore personal computers v/ ill be used in 
education activities, 'ihe first line 
sec up to na-~s produce conouters was 
designed to put out 20, CO./ per month. 
<itn three companies now delivering 
nachines, the really large (potential) 
Producers have not yet announced their 
jroducts. I expect nearly half a 
nil lion units to 09 solo in the first 
/ear, and in a few years tne nu;nber use J 
in education will be counted in the 
nil lions. .-'ineri the measures of expense 
and capability of a technology improve 
:)v two or three orders of magnitude the 
effects are more than jus t quantitative. 
Personal computers will brinq about 
qualitative changes in education in the 
nome an-i pernaps in colleges. 

Ihe seconj characteristic of 
personal computers is responsiveness. 
i'he nacnine is as available and responds 
as quickly as a personal and portable 
electric typewriter or television, iiut 
■lore than that, it responds effectively 
to :nore complex directives, rearranqino 
and processing information in n personal 
-vaynot possible with a shared system 
iesiqned for the average user. 
rurthermore, the high rate of transfer 
of data between memory, processor and 
display screen opens up new 
opportunities for real-time animations 
as well as data analysis. 

f;ie third aspect of the revolution I 
want to emphasize is personal control. 
The owner of a personal computer 
let ermines tne uses of one or more 

persona i machines, and snnges cheir 
cuaracter Lsti cs to personal needs an-j 
preferences. Perhaps the most important 
aspect or coutroi is tne intanoiole 
consideration of self-determination* 
the owner can liberal ly wrap his or ner 
arms around the machine, carry it about, 
paint it purple, arid love it 
as a pet, as well as re program it 
according to personal preferences. 

;e tailed- characteristics of personal 
computers arr' their current uses are 
receiving careful attention by CKLi 
staff at tne university of Pichigan. 
fheso notes attempt only to list a few 
characteristics and limitations. 

Li mi tat ions 

In 19/'/ tne limitations of personal 
computers may .have outwei:;neu their 
observed advantages for use within 
col le fie learning activities. However, 
19/.) brings more reliable and oowerful 
models of those inexpensive devices. 
Indeed, we can expect the capabilities 
to at least double each, year without an 
increase in production costs for at 
least tne next twenty years. 

Pa 1 i auil i ty has been a major problem 
vitn many nicrocompter systems to -Pate, 
do w tiiat personal computers are beinu 
mass produced, tne inteoraced systems 
perforin much more reliably tnan their 
predecessors which were usually built oy 
nobbiests from kits. However, the 
construction of peripherals for 
printing, storage ana the like does not 
measure up i? the same standard of 
reliability. ihe major problem faced ov 
manufacturers of peripherals is a need 
to or in ; out a very low-cost device, 
somethin.; comparable with the low cost 
of tne microcomputers themselves. iwe 
electromechanical devices for printing, 
drawing, or storing information on 
magnetic tape can not be both cheap and 
reliable. However, the prospects are 
good for reliable peripherals using new 
technical developments which will 
eliminate entirely the noving parts! 

Speed of processing is a 
consideration. Happily, .nost 
instructional applications involve 
simple programs and tne observed 
response from the personal computer is 
not distinguishable from that of a 
timesharing system. However, some 
applications require many computations 
(eg, for reducing data or analyzing 



BOX 1579, PALO ALTO CA 94302 

text), *i">j tne delays ma/ just be too 
great for a student interacting at the 
slow speeds of an inexpensive, personal 
compu ter. 

Personal computers do not now 
provide inexpensive access to large data 
oases, aiij updating tne Jat^ base i s an 
expensive operation if every user has an 
individual copy. Therefore, the use of 
personal computers in records handling, 
data reduction, ana shared files is not 
advisable with IvY/ technology. Tnis 
could cnange by the end of 19/d, 
no-// ever . 

gompacibili ty a;non.-i different 
personal computers imposes so:ne Units 
on exchange and mar ke ting of programs. 
Actually tins limitation is no more 
severe than for sharing pronrams among 
different computers made o/ the same 
vendor. dome of the reco emendations 
given in the third section of this paper 
will help remove this limitation. 

Probably tne standards that emerge 
will depend on certain features of the 
technology! inexpensive, read-onl/ 
storage; processors which are less 
expensive than tne medium- of curriculum 
storage; and audio-visual media wnich 
incorporate digital logic and storane. 

Approaches to applications 

■ dthin current liiiii tations of 
reliability, speed, storape, and 
compatioilit/ CtiL'i staff already find 
many applications for personal computers 
wfthrn the University of •Vichinan today. 
4a chines becoming available in the first 
half of iy/;3 will make valuable 
auditions to the resources for learning. 
Some machines acquired by individual 
faculty nemoers are being used in 
classroom demonstrations. At least 
three curriculum development projects 
are using personnl computers in 
connection with laboratories 
(preparation for regular lab sessions, 
information processing aids in carrying 
out the work, and data Generators for 
simulated laboratory activities). 
Libraries are considering use of 
microcompters in providing services on 
campus. ' Many individual students have 
purcnased machines, and some of these 
are oeing used for word processing 
(preparation of papers), computation, 
and information organization and 
retrieval. The School of '4usic is using 
a microcomputer system for music 
analysis and synthesis, and some 

else ironic music composition. 

joma tnou mt about the future of 
computing in education e.n-i society may 
ielo college teacners take a useful 
perspective oi~\ computer use o/ students 
tou.a'/. i liis topic has been written on 
at l«no-tn; i will list just three 
aspects i believe to oe very significant 
for educators considering the general 
tooic of computers in education: 
computer literacy, home entertainment, 
and comouter-basec videodiscs. 

\ oeneral literacy about computers 
is likely to be as widespread as 
knowledge of driving a car or skills for 
use of a liorary. Already electronic 
calculators nave oecome so inexpensive 
and straightforward to use that nearly 
anyone enn acquire a convenient device 
to" carry in a pocket, purse or wear on 
wrist. i'he impact has been primarily 
om of accomplishing calculations more 
reliably. As general information 
processing facilities become as common 
as tne portable home typewriter or 
television set (indeed, they will no 
douot oe incorporated within home 
tyoewriters and television sets), =) 
general literacy about their use will 
develop rapidly. To some extent tnis 
literacy is alreauy oeing accomplished 
in computer courses in intermediate 
scnools throughout the country, 
oel f-instruclion at home will take care 
of tne rest. All college students will 
kno,v now co A uu .iXPbUi 'i0 use comouters 

in is year, microprocessors- have 
oecome a significant part of the home 
entertainment industry. Computers are 
evident in video names which duplicate 
arcade devices, and in board games which 
play a fair game of chess or checkers or 
othello or backgammon. As 
general-purpose computers become 
incorporated in home entertainment 
centers, the capabilities for 
simulation, model line; and records 
handlini expand rapidly. An important 
oart of the nome market will be 
self-education, tfany companies will be 
selling learning packages which include 
computer programs to rur. on the home 
computer as well as supplementary 
materials and guidebooks. 

N-.5W developments in videodiscs nay 
revolutionalize personal computers, ror 
some years companies have been working 
on one version or another of the home 
videodisc. Jne of the driving forces 
behind this movement is the expected 



BOX 1579, PALO ALTO CA 94302 

market for old movies and special 
interest programs in the norne. However, 
in order to solve some of the technical 
proDiems of retrieving the bits of 
information which make up a video 
signal, the researchers turned to 
certain Digital techniques which are of 
considerable interest to computer 
designers. It appears now that 
including a microprocessor in the 
videodisc player will make it more 
practical, reliable, and economical. 
Computer control of the videodisc 
player provides an exciting resource for 
educational use of microcomputers. A 
small archive of movie clips (30 minutes 
total) or a very large archive of color 
still frames (54,000) or any combination 
can oe searcned and controlled by a 
personal computer. This means one can 
retrieve, display at regular speed, slow 
motion or still frame, and move to 
anotner section of the file based on 
instructions from the push-button 
controls and a history of interaction 
with the particular learner. 
Furthermore one can set up the personal 
computer to interpret the 54,000 frames 
of video as nearly a trillion bits of 
digital memory. This reduces 
consideraoly the present restrictions on 
microcomputers for handling large data 

These future prospects are quite 
exciting and prooably will bring 
significant and desired benefits to 
higher education and to society. A 
summary of trends concludes this 

Trends in computing, communication and 
the locus of education 

A number of changes in the 
technology of computing and 
communications, and changes in education 
and society, have increased professional 
interest in computers in learning and 
teaching, hour points are listed here: 
The low cost of microcomputers merit 
rethinking uses of technology in 
education. Inexpensive communications 
may shift the role of centralized 
educational computing. Incentives will 
improve for commercial production of 
materials for training and continuing 
education. Home computing, combined 
with other technology, will support a 
trend toward more education in the home. 

ihe low cost and easy use of 
microcomputers are providing access to 

1 '* j'Ia ; j.i I o ^ i o s^ x v " uvjiiipu i, j. : r.j iiiCl'jOllig 

graphics for many more people and 
projects. video techniques, such as the 
microcomputer-based videodisc player, 
are enhancing graphics and low-cost 
storage for instructional computing 
(Bork, Luehrmann pr\<i Schneider, in 
press). The economic and social 
pressures on educational institutions 
are forcing tie cis ions by state and 
federal agencies, and by individual 
institutions, to support additional use 
of technology. 

Inexpensive communications have oeen 
promised via satellites and optical 
fibre telephone lines. Very low cost 
and very large memories are proposed for 
educational time sharing system's. dhen 
these developments are realized, tne 
cost of centralized educational 
computing systems will drop 
significantly. i'his cost improvement 
will help restore some of the education 
market for timesharing taken over oy 
personal computers, but also it will 
increase the use of personal computers 
through low-cost support networks. 

Commercial incentives are emerging 
in tne U.S., now that packaoing is 
becoming more practical and the 
educational market is becoming 
interlinked with other computer uses, 
t-or example, recreational uses of 
personal or nome computers include some 
instruction for the very young learner. 
And professional uses by tne adult 
practitioner include some new 
opportunities for learning. New markets 
for training and continuing education 
nave focusea attention on the need for 
improved instructional desiqn. 

People who work on computers in 
education will have to pay attention to 
the dramatic development of personal 
computing. Already considerable 
education does take place in the home, 
and more money is spent on home 
education materials and correspondence 
in tne U.S. than on institutionalized 
education directly. I am sure a great 
deal of computer use in education and 
training in the next decade will be 
through inexpensive but rather capable 
machines purchased for personal use in 
the home and office. 

All tnese current developments can 
not oe covered in the pages of this 
article or in the corresponding session 
at the Faire. However, this article car 
be used together with others in the 



BOX 1579. PALO ALTO CA 94302 

Proceedings and tne suggested readings emitters in education ars essential to 

to expand your view of microcomputers success of new programs in public 

and uses of personal computing in nigher schools. 

education and to establish means to keeo Psrsonnl coi.iputino in homes shows 

in touch with a diverse and rapidly the greatest prospect for impact on 

changing fielu. learning about the computer. &/ the enc 

of IWd, personal computers //ill be in 

use in nearly half a million homes and 

aIAJS Or USti offices. lany of these devices may be 

used only with packaged application 

Learning about the computer programs, as are the preprogrammed video 

Learning about the computer is the games? but all tne equipment will be 

most rapidly growing area of computer programmable and include convenient 

assisted learning. It's use in CAL local storage for saving user-desianed 

follows quite naturally from the programs, 
introduction of computers into many 

parts of society and into homes Learn in j through the computer 
directly. As experience with computers In the past, the core of CAL 

becomes commonplace for a class of activity was dri 11 , practice, diagnostic 

learners, educators can use the testing, and question and answer 

computer and its procedures as metaphors tutorials. xoday these continue to be 

for' other processes and constructs. the major modes of learning in 

education about computers provides tools operational U.S. systems (e.g., the 

useful in other learning. The following PLAi\) 3 /stem of the University of 

paragraphs comment on computer literacy, Illinois and Control Jata Corporation, 

professional development, in-service Computer Curriculum Corporation systems, 

training of educators, and personal Mewle tt-Packard and Digital equipment 

computing. timesharing systems, and tne Univac 

Computer literacy for all students system), uven though these modes will 

is established now in some colleges. oe overshadowed for a time by a dramatic 

Computer programming may oe required (or growth of learning about and learning 

assumed and offered only as a non-credit with tne computer, they will remain 

course) so computing skills can be strong instances of computer 

assumed in mathematics, sciences, application, readily accepted because of 

engineering and some other professions. their familiarity, moderate cost, and 

the extent -of c-^mputer literacy convenience of to.tal .systems. 

throughout tne U.S. is still very low, 

out inexpensive computing devices and Learning with the computer 

the popularity of the activity with ihe computer as an aid to learning, 

students will increase the percentage and an adjunct tool for tne learner, is 

dramatically in the next few years. taking on new dimensions now that 

Professional development regarding computing is inexpensive and portable, 

computer use is common for many areas Hand calculators used in the laboratory, 

already; accountants, bankers, engineers classroom, and study hall have taken on 

and others whose professional activities the characteristics of computers* 

depend on automatic computation and stored programs, program libraries, 

information processing are refreshing peripheral storage, printers, graphics, 

skills and ootainina new ones. etc. As these least expensive 

Additional professions are finding computing devices take on general 

computer assistance indispensible, and characteristics, the general-purpose 

information is offered through special computers are decreasing in price to 

courses and institutes for architecture, match the calculators. Practical uses 

law, medicine, and others. include simulation, gaming, problem 

In-service training of teachers is solving, and creative activities, 
essential if public education is to Simulation and gaming have always 

catch up with the rapid change of been popular with teachers and students, 

technology. University institutes offer .tow these entertaining activities are 

skills training during the school year appearing in academically respectable 

and in the summer. Sound information textoooks and laboratory materials, 
and constructive attitudes regarding Problem solving activities once 


required knowledge of programming, 
jnless cleverly imbedded in a tutorial 
sequence which prompted for the 
oarameters and equations or whatever was 
needed. Now problem-oriented languages, 
and familiarity with keypress sequences 
Dn programmable calculators, put problem 
solving in reach of any learner 
experienced in the discipline. Students 
can be asked, to turn to a problem 
solving facility on the CAL system (or a 
calculator at hand) to carry out some 
computation or modelling activity. 

Creative activities in education 
are aided by computers very 
"licely, at least in experimental 
systems. Perhaps the most impressive 
Dverall is the first approximation to 
the •• aynabook" developed by Alan Kay and 
nis colleagues (ly//). Young children 
are able to sketch, animate, compose 
Tiusic, arrange words in poetic forms, 
and carry on many other creative 
activities usually reserved for 
specialized and advanced users of 

Learning support systems 

Computer managed learning is 
expanding within" the U.S., albeit 
quietly and often without any note of 
the computer role. A significant 
percentage of schools are using computer 
systems to aid in classroom management. 

Information management goes beyond 
instructional management to help the 
student as well as the instructor with 
information needs. Guidance systems 
have become quite popular, including 
SIGI of educational Testing Service. 

The automatic generation of learning 
and testing materials by computer will 
soon become a common activity among 
computer aids. Some tools are already 
q~uite popular, especially computer 
assisted test generation (Lippey, 19/4). 

Some pointers to current work in the 
United States have been provided within 
the text. The next section provides a 
sample of applications in various areas 
of instruction. 

Areas of instruction 

The broad range of computer uses can 
be shown by selecting some of the less 
likely uses in six areas of instruction: 
math, sciences, social sciences, arts 
and humanities, languages and 
communications, and the professions. 

The few instances given here represent 
only a small part of all the computer 
aided learning curriculum materials. 

Students in a mathematics course 
nave used a simple computer Inn gauge 
(LOGO) to generate a mathematical systex 
building from primitive elements. 

In a laboratory course in chemistry, 
preparation for use of titration 
equipment is aideu by conceptual 
experience provided economically to 
individual students who us* the graphic 
animation capabilities of the PLATO 
System before they go into the 

A simulated laboratory provides 
research experience for undergraduate 
students in psycnolocy. Tne computer is 
used as a data generator. The value of 
the simulation depends on the activity 
of a classroom research community and 
the effectiveness of the teacher as a 

In the arts ana humanities computing 
serves as a medium of communication as 
well as a tool for creative work as part 
of learning. Perhaps the most 
interesting instructional use in 
humanities today is as an aid for 
scholarly work by the student. Graduate 
students of literature have used 
preprogrammed applications packages in 
exercises to determine authorship or 
analyze style. Undergraduate students 
explore rules of language through 
computer generation of poems and 
stori es. 

In language learning, aid in 
practice of skills is the dominant 
computer use. Contrary to the idea that 
the cost of development needs to be 
distributed over many students, two 
professors at Stanford University are 
programming computer assistance for a 
dozen specialized courses which have 
such low attendance that the department 
of Slavic languages can not afford to 
staff them. With this assistance 
(tutoring, drills, and practice 
exercises) the professors plan to .handle 
a larger numoer of students and in a 
greater variety of courses than 
previously possible. 

Preparation for professional work 
accomodates as much computer use in 
training and education as anywhere. Foi 
example, management games are very 
popular in business education and 
natural resources; simulated cases are 
used in lav; and medicine; and design 



BOX 1579, PALO ALTO CA 94302 


exercises depending on computer 
assistance are common in enaineerin- 
architecture. One of the most 
innovative applications is computer 
assistance for advanced seminars which 
bring together graduate students from 
different departments for study of a 
proolem area (e.g., energy conservation, 
regional planning, or technology 
assessment), .each participant uses 
computer assistance for organizing 
Ihfof nation from diverse and sometimes 
unfamiliar areas, communicating with 
others in the seminar between 
face-to-face meetings, and drafting 
working papers for review by the group, 
i'he organizers of the seminar keep the 
group focused on the problem without 
minimizing important background 
material, and call on resource people 
who might not otherwise have time to 
participate except for the convenience 
offered oy computer-assisted 
conferencing. This enables them to 
respond at any time of day, any day of 
the week, and from any user terminal 
which can connect to the computer or 
network handling the conference. 

.MciiJHU L/£VHLOPMi5h1'3 

^eeds and issues 

Further development of computer 
assisted learning in the U.S. will 
interact with a number of needs and 
i issues re la ted to the te chhoTbgy as we'll 
as to educational applications. Three 
such issues are listed here. 

Microelectronics technology, with 
rapidly decreasing costs which depend on 
greatly expanded usage, is forcing 
producers to find new markets for 
computers and related technology in 
education as well as throughout society. 
Appropriate uses in education require 
planning for and managing the design of 
the technological aias and their 
introduction into educational activities 
and institutions. Personal computing is 
being marketed strongly in the U.S. and 
will be taken up in the next few years 
by many people for entertainment and 
small business purposes. The equipment 
in which industry is investing large 
sums of money for the personal computer 
market will not serve educational 
purposes well without attention to 
considerations of design specifically 
for educational purposes. 

dinner education faces serious 
problems of financing, access, 
credibility, and the like. Certain of 
these difficulties can in part be ensed 
oy appropriate uses of technology, if 
resources are available at the right 
time and place for research, 
development, evaluation, demonstration, 
diffusion, and operation (Levien, 19/2). 
Otherwise, significant, opportunities to 
aid' all learners, and in particular the 
disadvantages, the handicapped, the 
gifted, and the isolated learners, -nil 
be lost. 

Jeneral literacy in computing and 
information systems will oe required by 
all, consumers as well as marketers, 
employees as well as managers, and 
learners as well as scholars, if societ; 
is not to be disrupted by a revolution 
encouraged oy rapid growth of technolog ; 
needed to support today's "information 
society." Computer literacy can begin 
in elementary school, arid sooner. 
However, oecause of the rapid 
introduction of the technology in many 
areas, computer literacy training must 
oe carried on in colleges, professional 
schools, certification programs, 
on-the-job training, community 
education, and public meoia reaching th« 

nomes . 

Of course, otner problems and issues 
may oe at least as important as the 
three listed above. However, this 
summary statement does express the 

nature of the current si ttia-t ion for 

computer assisted learning in higher 

Areas for possible action 

One major need at the national level 
is for coordination of planning and 
funding, taany important matters, e.g., 
goals, standards, and credibility, can 
be accomplished only through national 
discussion and action. fhe funding 
necessary to meet the needs in this are; 
can no longer be handled piecemeal 
through many different agencies. The 
commitment to excellence in education 
and to effective use of technology must 
come from the top. Information systems 
in education and society cut across man- 
areas* research, development, 
handicaoped, gifted, elementary, and 
professional. Useful information and 
good advice on these matters has been 
accumulating for over 15 years in the 
form of recommendations of national 



BOX 1579, PALO ALTO CA 94302 

commissions (National Academy of 
Sciences, 1966, PSAC, 1967, Commission on 
Instructional Technology, 1970, Carnegie 
Domxiission, 1972), professional 
organizations (e.g., CBMS, 1972) and 
review projects (Zinn, 1970, Anastasio 
ana Morgan, 1972, Hamolen, 1972, Levi en, 
19 72, Mosmann, l97o). A new commission 
or conference can begin with the 
recommendations of earlier efforts, 
review the present state: of technology 
3nd institutions, and take thoughtful 
action. Action is necessary now in 
response to the pressures and problems; 
furthermore, benefits are more easily 
justified today in light of the 
dramatically improved economics for 
applications of microelectronics and 

Designated centers with good support 
could provide sites of excellent 
training, development and research 
(Camegie Commission, 19/2). Potential 
jsers urgently need the most current 
information, the best training, and an 
optimal environment for development 
activity. Residencies would provide an 
opportunity for individuals to get away 
from regular responsibilities to 
initiate new work. Research 
opportunities would be increased 
significantly by bringing together 
creative individuals with necessary 
resources and a variety of learning 
environments. Large-scale exploration 
of technological opportunities and basic 
concepts of learning would become 
oractical. Alternative CAL systems and 
approaches to curriculum could be 
compared within the same environment. 

Immediate action to give educational 
jses for computers an identity different 
from data processing would in some 
organizations facilitate obtaining 
equipment necessary for meeting 
institutional goals at lesser costs. 
Instances include public education as 
A/ell as military training, and state and 
local support as well as federal. 
Immediate action to recognize 
computing and information processing as 
a significant part of basic education 
tfould set in motion the process of 
curricular revision necessary to the 
information age. In a few years all 
students in public schools would become 
familiar with computers, programming as 
tfell as applications, by about the 
eighth year of school. The college 
teacher could then assume long 

familiarity witn computers for entering 

Curriculum development requires 
special attention, since computer use in 
education is a new industry, as yet 
untested and lackinn incentives for 
developers and distrioutors. Commercial 
publishers can not be expected to 
initiate nign risk ventures, and yet 
they may be left behind if the computer 
vendors try to provide curriculum 
materials. Universities 3nd colleges 
have much to contribute since most 
textoooks originate there today. 
Individual authors need to see rewards, 
ooth academic and economic, for their 

The social implications of dramatic 
changes in availability of information 
and automatic processing require 
attention, hvery elaborate clipboard or 
binder will nave a small pocket for a 
specialized calculator, fiver/ reference 
book or procedures guide will have 
imbedded within its cover an infor nation 
processor suited to trie subject. each 
desk encyclopedia will include sounds 
and animations and a microprocessor 
which conducts searches of the entire 
text as well as selects from the 
contents, index and cross references. 
Planners need to consider the 
implications of new modes of 
representation and communication with 
macnines, new skills for learning, 
problem solving and creative activities, 
new roles for teachers and managers, ne'< 
situations for learning at home, on the 
joo, and in the community. 
Administrators need to plan systems so 
that increased dependency on informatior 
machines, for assessment of ideas as 
well as retrieval of information, will 
not become an inappropriate crutch. 

Communication between people and 
machines needs careful attention. As 
long as the students (or other casual 
users) need to type on a keyboard and 
watch for text and numbers and simple 
diagrams to appear on a special screen, 
these machines will have a rather narrow 
application in training and education. 
However, when a user can talk to the 
machine and get a response not just in 
printed text but in spoken words and 
other sounds, and can see the effects of 
his or her directives in the actions of 
equipment such as models and tools, ther 
the computer will fit in to a much 
larger world of learninc 



BOX 1579. PALO ALTO CA 94302 

Additional areas for possible action 
at least as important as tne eight 
listed above can be added, nor example, 
through computers special onportuni ties 
become available for the ufted student. 
Dramatic improvements in communication 
and learning are achieved for the 
handicapped: Kurzweil's reading machine 
and ielsensory's talkinr calculator and 
Braile output device for the blind, and 
similar specialized equipment for the 

Other aspects of the technology 
provide facility for producing speech, 
processing knowledge, and building 
personal skills for learninq and 
performance . 


/■tew modes of representation and 

Future aevelopnients will extend the 
modes of communication possible between 
the learner and the machines. Speech 
and other sounds will become suitable 
for entry into the machine as well as 
for output from it. Gesture and other 
motions will be interpreted usefully. 
These developments will bring immediate 
benefits for' those lacking some standard 
sense or accepted communication means. 

for. e_xample , computer-ass is ted 

communication will revolutionize Braile 
for the sightless, and provide random 
access to audio and other personal 
notes. Already an application of 
microcomputers provides speech to those 
who have lost it through accident or 
cereoral palsey. Physiological measures 
will be encorporated as input, opening 
up new channels for persons lacking the 
motor control necessary to operate 
typewriters or to speak. 

Information will be represented with 
improved graphic and auoio means using 
networks and other data structures. The 
exploration of knowledge will be more 
directly available through manipulation 
of structure, organization, and dynamic 
interactions by the learner. The 
student will, with assistance of 
information processing routines, work 
effectively within a personalized and 
dynamic information base. This 
development will depend on considerable 
literacy in infomation handing using 

iJew skills of learning, problem solving 
and creative production 

Computer assistance will help 
learners arrange multiple views of text 
and graphics. "Skills of speed reading 
will be extended by aids for i 'mediate 
comparisons and cross references among 
sections of text. Facilic/ with 
graphics will extend far beyond the 
multi-media shows of tocay. Authors of 
textbooks and reference materials face 
new challenges in apply in? the 
technology and anticipating improved 
skills of users. 

Problem solving skills will expand 
in more directions than can be 
anticipated. Induction may be 
facilitated oy interactive, 
computer-assisted deduction. Proof of 
the four-color map problen by students 
with computer assistance is only a 
beginning. More creative solutions to 
engineering problems will be especially 

Artistic creations will similiarly 
be extended beyond our present abilities 
to comprehend ana appreciate, r'or a 
primitive example, consider today's 
dynamic sculpture for which a 
sound-light score interacts with tne 
movements and speech of its observers. 
Young artists will find many new 
opportunities, and the world of creative 
art will be opened to those previously 
excluded by physical handicaps. 

Previously untapped capacities for _ 
learning and performance w'i 11 be put to" 
use. A computer-based lab in whic.i 
learners explore their own abilities may 
help in areas such as inter-hemispheric 
communication ("using both sides of the 
brain") and enhanced mechanisms of 
recall and pattern recognition. 
Unanticipated and dramatic benefits may 
follow from the development of synergism 
of the human user and various machine 
information systems. Some of the most 
significant benefits may be obtained 
with the enhancement of communication 
within communities of all sizes. 

The United States does not have the 
only innovative projects working in 
these areas, as is apparent when 
reading other articles reporting on 
developments elsewhere in the world, 
extending communication, processing 
knowledge and building personal and 
interpersonal skills are important areas 
for future development anywhere in the 
world of personal computing and learning 
today . 



BOX 1579, PALO ALTO CA 94302 


Anastasio, Ernest J. ana Morgan, Judith 
S. (1972) Factors Inhibitino' tne Use of 
Computers in Instruction. Educom, 
Princeton, N.J., USA. 

i31urn, Honala (Ed). (19/1) Comoutprs in 
Undergraduate Science Education. 
Conference Proceedings, Commission on 
College Physics, College Park, Mar/land, 

Bork, Alfred M., Luehrmann, Arthur u . 
and Schneider, Edward. (in press) 
Report of a Conference on Intelligent 
Videodiscs. University of California, 
Irvine, USA. 

iirignam, Christopher !?., .<amp, Martin 
and Cross, Kenneth J. (1975) Index of 
CAI ..iedical Education, Lister Hill 
Center, National Library of Medicine 
Bethesda, Maryland, USA'. 

Carnegie Commission on Fisher 
Education. (1972) The Fourth devolution* 
Instructional iechnology in Higher 
Education. McGraw-Hill', New York, USA. 

Commission on Instructional Technology. 
(19/0) To Improve Learning* An Eva uat ion 
of Instructional Technology, r. h. 
Bowker Co., New York, USA.. 

CONDUIT. (19/3) Documentation 
Guidelines. CONDUIT, P.O. Box 388, 
Iowa City, Iowa, USA. 

CONDUIT. (19//) Pipeline. CONDUIT, 
P.O. Box 388, Iowa City, Iowa, USA. 

CCUC (19 7/) Proceedings of the 19 7/ 
Conference on Computers in the 
Undergraduate Curricula. Michigan State 
Unmversity. CCUC, 1243 Linquist 
Center, Iowa City, Iowa, USA. 

Hamblen, John. (1972) Inventory of 
Computers in U.S. Higher Education, 
1969-70, U.S. Government ""PrTnEinq 
Office, Washington, D.C., USA. 

Harrison, Shelly A. and La.vrencp I 
atolurow (Eds). (1974) Educational 
iecnnologies* Productivity in Higher 
education. State University of N^v 
fork, Stony .irook. New York, USA. 

Hunter, Beverly, Kastner, Carol 3., 
Kubin, Martin L., and Seidel, Robert J. 
M97o) Learning Alternatives in U.S. 
education* wnere Stuaent and Computer 
:.ieet, educational Technology 
Publications, Eng lewcod Cliffs ,>■* 
Jersey, USA. 

Kamp Martin and brigham, Christopher 
(bds.) (ly/3) special issue of 
Computers in Biology and 4edicine. 

Kay, Allen, and Adele Golaberg. (19//) 
Personal Dynamic Media. Computer. 10 
3, 31-41. 

Levien, rfoger E. (19/2) ine Snernina 
iechnology* Instructional Uses of'th* 
Computer in Higher Education, A Carnegi- 
commission on Higher Education and >ianV 
Corporation Stud/, McGraw-Hill Book- 
Company, New York, NY, USA. 

Lippey, G. (Ed) 6 i974) 
Computer-assisted Test Construction, 
educational iechnology Publications, 
englewooo Cliffs, NJ, USA. 

Luehrmann, Arthur W. (19/2) Should the 
Computer Teach the Student or Vice 
Versa." Proceedings of the Sprino Joint 
computer Conference. AMPS Press, 
Montvale, New Jersey, USA. 

Mam, uana 3. (19/6) Experiment 
ovulation (cXPEfi SIM)* The Development 
of a Future-Oriented Pedagogy, in D. 
Bailey (to) , Computer Science in tie 
Behavioral and Social Sciences. 
University of Colorado, Boulder, 
Colorado, USA. 



BOX 1579. PALO ALTO CA 94302 

.iorton, A. Kent and Luahrnann, Arthur 
d. Jr , (W/'j) Project CO.fPUTe: A 
Mechanism for Producing and Distributing 
Instructional Material. ^3--W7 in 
Computers in education, 0. Lecarme and 
R. Lewis, hds. mrth-Ho Hand, 
\ms terdam. 

/,osmann, Charles. (lV/o) e'valuating 
Instructional Computing: -'.ensuring Needs 
and Resources for Computing in Miqner 
education. The University of 
California, Irvine, USA. 

iJational Academy of Sciences. U9o6) 
Digital computer Needs in Universities 
and Colleges, A Report of the Committee 
on Uses of Computers. National Research 
Council, i'.asninqton, D.C., USA. 

Nelson, Ted. ihe Home Computer 
devolution. Published by the author. ^ 
distributed oy The distributors, 702 5. 
Michigan, South Send, lh 466 IB. 

devison, John ./,. (1976) Computing in 
the Liberal Arts College. Science. 
194, 39o-402, October. 

Nievergelt, Jurg. (W/d) Interactive 
Systems for education: 'ihe New Look of 
CAI . Proceedings of tne Ir'IP 2nd .'lor Id 

Conference ..on. .Computer. Ldu cation, _. 

Amsterdam* dorth-Ho] land. 

PSAC. (Wo/) Computers in Hiqher 
education. President's Science Advisory 
Committee, U.S. Covernrrent Printing 
Office, .Jasninqton, U.C., USA. 

Seidel, Robert (e'd). (1975) 
Proceedings of the Ten-Year Forecast for 
Com outers" and Communications. Human 
Resources Research Organisation, 
Alexandria, Virginia, USA. 

Seidel, Robert J. and Hunter, oeverly 
C. (co-inventigators) (1977) Academic 
Comouting directory: A Search for 
exemplary Institutions Usinc Computers 
for Learning and Teaching. (First 
edition) Human Resources Research 
Organization, Alexandria, Virginia, USA. 

Suooes, Patrick. (I97S) I.npact of 
Comouters on Curriculum in the Schools 
and' Universities, in 0. Lecarme and H. 
Lewis (ws.), Ii-IP Second dorld 
Conference on Computers in education, 
Part 1. lOrth-Kollarvi Publishing 
Company, 1 1 3- I /'9. 

Suooes, Patrick. Smith, Rooert and 
! J ear J, Marian. (1977) University- level 
Comouter-assisted Instruction at 
Stanford: 1975. Instructional Science, 
6 , I 5 1 - 1 cj5 . 

■/an Jam, Andries. (W/o) Computers in 
Teaa-iing: An Application of Hypertext. 
Final Report to .national endowment for 
tne Humanities, drown University, 
Providence, Rhode Island, USA. 

.Shite, dames. (19//) /our Home 

Computer, dY/'.AX, 

Menlo Park, California, USA. 

Zinn, i'.arl L. (W/U) An evaluative 
Review of Uses of Computers in 
Instruction (Project CLUe» Computer 
Learning under evaluation). Final 
Report U.S.O.e. contract no. 
OeC-o-9-32-509-0032, University of 
Michigan, Ann Arbor, Uchiian, USA. 

Zinn, Karl L. (1977a) Free and 
Inexpensive Materials oh Computing in 
reaching and Learning Activities: An 
Informal Appraisal. Center for Researc) 
on Learning and Tea chine, University of 
Uchican, Ann Abor, Michigan, USA. 

Zinn, Karl L. (1977b) Computer 
Facilitation of Communication within 
Professional Communities. Behavioral 
Research .-.iethods and Instrumentation, 9, 
2, 95-10/. 

Zinn, Karl L. Computer Assisted 
Learning in the United States. In a 
special issue of the British Journal on 
Programming and educational Technology, 
in press for spring /d. 



BOX 1579, PALO ALTO CA 94302 


Thomas A. Dwyer 

University of Pittsburgh 

Pittsburgh, PA 15260 

Introducti on 

Is there a better way? If so, 
what is it, why will it be better, 
and is there assurance that this 
time we can "get it right"? This 
paper argues that the answers are 
to be found in personal computers, 
and that the right way to use them 
is staring us in the face. To see 
why, let's examine a remarkable 
learning experience many readers 
have had. 

The Secret Unveiled 

Imagine you have just arrived 
at an airport in a strange city, 
and now need to reach your final 
destination by way of unfamiliar 
roads. One possibility is to take 
a taxi. This option is direct and 
efficient, and it can have the 
bonus of being a personalized tour 
along a tried and proven route. It 
has all the potential for being a 
first-rate educational experience. 
Yet, it is not likely that upon 
arrival you could pass a test asking 
for an accurate description of the 
route just taken. Your "individu- 
alized" treatment will have gotten 

you to your destination, but you'll 
still be a stranger to the territory. 
The ingredients for a true adventure 
were missing. 

Consider now another option. 
Suppose that you rent a car, and 
drive yourself. The simple act of 
moving into the driver's seat will 
have a profound effect upon the hun- 
dreds of interactions about to take 
place as you move into the role of 
problem-solver, becoming an adven- 
turesome and necessarily creative 
learner. This option comes at a 
price, of course. There will be the 
need to find and negotiate for a car, 
ask directions, study a map, choose 
between alternatives. There are also 
likely to be mistakes -- "ineffi- 
ciencies" by some standards. Land- 
marks missed, or instructions mis- 
understood will mean backtracking 
and revised planning. Questions will 
have to be asked, time will be lost. 
And the cost of having exclusive use 
of a car will be higher. But in the 
end, the person who goes "solo" will 
have learned things aboutgetting 
from A to B that are accessible in no 
other way. 

The paradox we see from this ill- 
ustration is this: the guidance of 
others may very well inhibit the best 
kinds of human learning. The con- 
clusion it suggests is that people 
have far more instrinsic talent for 
the business of learning than they 
have for the business of describing 
it,' or bringing it about in others. 
Even the ability of a small child to 
learn to deal with the incredible 
complexities of a constantly changing 
world makes our ability to explain 
how it all works seem pale by com- 

Even less impressive is our 
success in promoting human learning 



BOX 1579. PALO ALTO CA 94302 

through institutions organized 
expressly for that purpose. 
Placing students within the walls 
of carefully designed schools 
would certainly appear to be the 
best way of assuring that they 
get transported along paths 
(called curricula) that visit 
many important educational points. 
The predicament faced is that for 
most students these are only 
visits, and dimly remembered ones 
at that. They never get to really 
know the territory. 

Enter the Computer 

What has all this to do with 
computers? The answer is both 
simple and perplexing. In 
reviewing the history o'f computers 
in education one finds that the 
majority of effort (and money) 
has gone to promoting their use 
as expensive educational "taxi- 
cabs." This effort has gone 
under the name of Computer 
Assisted Instruction, or simply 

are v 
be an 
of le 
r o le 
c i a r i 
of ed 
real i 
g i n a t 
ti ce. 
Run , 

c 1 u s i 
i ntro 
prof o 
was n 

n CAI , d 
iewed as 
the "dri 

sson des 
ammers , 
nts are 
of i-n-d-iv- 

is, of 
they wil 
es of pe 
portati o 
u c a t i o n a 
ty, the 

more 1 i 
i ti ous v 
i ve 1 and 
The al 
ny Oetti 

ri ver 

ver" t 
nal co 
and ha 
rel ega 
■i-d-u-a 1 
1 be t 
n to t 
1 worl 
promi s 
ke a s 
i s i t s 
s of d 
arm ra 
nger ' s 
Run , 

on that 
duce tec 
e d u c a t i o 
undly ig 
o t w i t h o 

n has 
ut bas 

and passe 
able e n t i 
urning ou 
, compute 
rdware ve 
ted to th 

the prom 
he benefi 
, customi 
he most e 
ds. In 
ed tour i 
huttle, w 
to the un 
rill and 
ised in 

1969 boo 
with its 

effort t 
i c a 1 chan 
reveal ed 

we still 
i s . 



t to 







xoti c 

s at 






are , 

Enter the Personal Computer 

But there is also good news 
and it's getting better all the 
time. It comes partly in the 

form of a new approach to technology 
and partly as a growing body of 
examples of student accomplishment 
that indicate an exceptional idea 
is at hand. It's a flowering of the 
"dri ve-yoursel f " option (called 
"solo-mode computing") in schools, 
made possible by personal computers. 

The new spirit is all the more 

remarkable in that it pretty well made 

it on its own. The funneling of 

large funds into CAI made early solo- 

mode computing an uphill battle for 

all but the most determined and adven- 

turesome. It is to the credit of the 

ingenuity of exceptional teachers and 

their students, that the door to solo- 

mode computing was gradually inched 

open. What first came through the 

crevice was at times amateurish, at 

times quite professional. But it was 

always inspirational and exciting, 

and it caught on. 

And now a new force from a com- 

pletely unexpected quarter promises 

to swing the door wide open. It's in 

the form of a personal computing move- 

ment that looks upon computing as a 

personally desired and appreciated 

enterprise. It's a use of technology 

that is (to use the words coined by 

Ivan Illich in his book De-Schoolinq 

Society) "convivial" rather than 

"manipulative" vis-a-vis the aspir- 

ations of people. It's the "drive- 

yourself" option for Gomputirvg -tome------ 


The personal computing movement 
is important because it can bring a 
new and different use of technology 
into our schools without major expen- 
ditures. If it finds a mixed welcome 
at first, there need not be concern. 
This is because it is a use of tech- 
nology that can literally exist in a 
home, a storefront, a community center, 
the corner of a library or museum. 

Actually ther 
sonal computing vn 
i n school s , becaus 
hands of the peopl 
pay for them, and 
It will thus be th 
of technology to e 
not sell itself as 
tionary." It guar 
Yet it assures eve 
in the right hands 
know the value of 

e is hope that per- 
JJ_ f i nd a wel come 
e it is in the 
e who use schools, 
care about them. 
e first application 
ducation that need 

"new and revolu- 
antees nothing, 
rything, because it's 
, free spirits who 
learning on one's own 



BOX 1579, PALO ALTO CA 94302 


Peter S. Grimes 

Curriculum Supervisor 

San Jose Unified School District 

1605 Park Avenue 

San Jose, CA 95126 

(408) 998-6124 

The purpose of this paper is to apprise 
the reader of how San Jose Unified School Dis- 
trict is utilizing personal and home type mi- 
crocomputer systems in its instructional pro- 
gram. San Jose Unified is considered to be 
somewhat of a bellweather in the educational 
use of microcomputers because it established 
a district policy for the rapid introduction 
of microcomputers into its secondary schools 
over two years ago when few educators knew 
the possibility even existed. The paper ad- 
dresses the following topics : 

• A general description of S.J.U.S.D.'s 

instructional use of microcomputers . 
•A brief rationale for the development 
of S.J.U.S.D.'s microcomputer policy. 

• An explanation of why we promoted in- 

structional microcomputing before 
identifying a curriculum. 

•A more detailed look at how we are us- 
ing microcomputers . 

•A brief exposition of some of the things 
we have learned. 

San Jose Unified School District is a 
large (38,000 ADA) urban school district 
nestled in the center of the great Santa 
Clara Valley at the southern end of San Fran- 
cisco Bay. We have seven high schools, seven 
junior high schools, thirty-seven elementary 
schools, and a regional vocational center. 

Our educational computing facilities in- 
clude ten time-share terminals (PDP 8/E), fif- 
teen microcomputers (one SOL, three POLY 88, 
one CROMEMCO Z2-D, eight IMSAI 8080 and one 
PET), and a Hewlett Packard 2000 system at 
our vocational center. (Since the program at 
the vocational center is dedicated almost ex- 
clusively to vocational data processing, the 
following remarks refer to our microcomputers and 
PDP 8/E time-share mini.) These terminals are 
generally distributed as follows: one or two 
terminals in each junior high, two or three 
terminals in each senior high and one roving 
terminal for our elementary extended learning 

Our use of these terminals is presently 
organized around the following themes . For 
junior high school, the emphasis is computer 
literacy for a large number of students, with 
some schools offering BASIC programming as an 
elective. The senior high school objective 
is computer programming using a variety of 
machines: programmable pocket calculators, 
desk top printing calculators (with optical 
card readers) and general purpose microcom- 
puters. Our language of choice is BASIC with 
the probable addition of disk FORTRAN in the 
near future. We also anticipate the develop- 
ment of an introductory computer science elec- 
tive as the need develops. 

We are told that San Jose Unified is some- 
what unique in its district wide policy to en- 
courage instructional computing through the 
purchase of small microcomputer systems. The 
dramatic thrust of this policy is evidenced by 
our purchase of fifteen such units in the past 
two years and a probable similar expansion over 
the next several years. As a school district, 
we are quite serious about promoting general 
purpose computing through small microprocessor 
based systems. 

I believe it is important to reflect upon 
the origin of this policy. Well over two years 
ago, it became quite apparent to us (residing 
as we do in the heartland of the digital elec- 
tronics industry) that small general purpose 
computing systems would become relatively in- 
expensive and thus well within the budget cap- 
abilities of individual schools and school sys- 
tems. (This judgement has been thoroughly vin- 
dicated with the marketing of the $595 PET in 
late 1977.) This belief led us inexorably to 
certain conclusions: 

1) Inexpensive computing would rapidly 
lead to a demand for the inclusion of 
computer programming and computer 
science into the public school curri- 

2) The ordinary citizen would very soon 
have to have some knowledge and skill 
in the use of many types of computing 
devices; i.e., pocket calculators, 




BOX 1579. PALO ALTO CA 94302 

programmable calculators and compu- 

3) Small general purpose computers would 
rapidly become important as instruc- 
tional devices (learning games, simu- 
lations, tutorials, drill and prac- 
tice) . 

4) Profound changes in the mathematics 
curriculum would become inevitable; 
i.e., increased stress on decimal no- 
tation, flow charting, algorithmic 
programming, iterative and recursive 
techniques, computer evaluation of 
functions, the demise of the slide rule, 
and vastly increased use of mathema- 
tical processes heretofore requiring 
calculations far too complex and time 
consuming for common use. 

In our mind, the digital electronics re- 
volution had a certain inevitableness about it. 
These things would happen! They would happen 
quickly! The sooner we became involved, the 
more control we would have over the situation! 
We installed our first microcomputer (an IMSAI 
8080) in March of 1976. 

San Jose Unified has become somewhat of a 
bellweather for schools and districts contem- 
plating their first entry into educational mi- 
crocomputing. We have received many inquiries 
about what we are doing with our microcomputers. 
These inquiries have a common element. What is 
your curriculum? What was your planning? What 
do students do with the machines? There seems 
to be a feeling that to justify the investment, 
the use of the machines must be totally maxi- 
mized within the shortest period of time. It 
is our opinion that this seeming urge to docu- 
ment the need for computer resources and spec- 
ify a curriculum is premature at this early 
stage. That is why we do not as yet have an 
identified curriculum. Our belief in the in- 
evitability of educational computing does not 
imply that we have gazed into the crystal ball 
and have thereby seen the future. We really 
don't know exactly how these microcomputers 
will ultimately be used in our schools. Nor 
does anyone else! All we know is that we must 
teach more about computers to more and more 
students and that we need computers to dis- 
cover how to do this . We think that teaching 
BASIC programming is a relatively safe thing to 
do at the present time. 

Why do we say that detailed early planning 
is premature? Because very few of us in sec- 
ondary public education have any real exper- 
ience with computing, even mathematics teachers. 
We have gone back to school ourselves! We are 
learning BASIC and FORTRAN, assembly program- 
ming, and something about computer science in 
general. We are finding out more specifically 
how our mathematics courses will probably be 

impacted by the advent of inexpensive cal- 
culators and computers. We are also finding 
out what the present student demand is for 
computer elect ives so that we will have a 
basis for projecting the future. We are try- 
ing to be creative. We are fearful of making 
detailed plans before we are knowledgeable 
enough to do so. We feel very strongly that 
teachers should develop the curriculum and 
that they would not be motivated to do so 
without the actual physical presence of a 
computer. (Teachers, too, can be very prag- 
matic. It doesn't make very much sense to 
make a large investment of after work time 
and effort unless the resulting skills and 
knowledge can be put to use . ) What we have 
done, then, in San Jose Unified is to provide 
the computer resources so that teachers could 
make the necessary discoveries and obtain the 
necessary background to begin identifying and 
meeting the needs created by the computer 
revolution. This development takes time. 
That is why we became involved so early. And 
so... what are we doing and what have we 

We have introduced programming courses 
in high school. These vary from one to three 
semesters. A typical sequence would be, a) 
introduction to computing using programmable 
hand held calculators and printing desk top 
units, b) BASIC programming with microcom- 
puters, and c) advanced BASIC programming. 
In junior high school we have taken three 
approaches. First, a few schools are offer- 
ing a one semester BASIC programming elective 
for grades 7-8-9. Second, after school com- 
puter clubs have been organized in which 
BASIC and system programming is taught in a 
more recreational context. Third, we are 
developing a plan for computer literacy where 
most students prepare a program, enter it into 
the computer and successfully run it. At the 
elementary level we are testing the waters 
with one roving microcomputer serving our 
grade 4-5-6 extended learning program (MSM) 
students. So far these students have been 
exposed to a variety of learning games, sim- 
ulations and drill and practice routines. We 
have also taught the more motivated elementary 
student some rudiments of BASIC. 

We have also learned a good deal. There 
is definitely a need for computer instruction. 
Five of our high schools are offering one or 
more semesters of BASIC programming. One 
high school has employed a computer science 
teacher and anticipates that he will be teach- 
ing full time in this area within the year. 
Although we have found no suitable texts, a 
large amount of reference material exists 
which has made it fairly simple for teachers 
to assemble a body of graded programming 



BOX 1579. PALO ALTO CA 94302 

experiences (mostly in mathematics). Up to 
300 high school students are currently involved, 
a number likely to grow as our computer re- 
sources expand and as computer programming be- 
comes more and more recommended or required for 
college matriculation. 

Our experience in junior high has been 
quite similar. Although there is not as great 
a demand for programming as an elective, the 
computer clubs have been extremely popular and 
wide scale computer literacy is being developed 
as a goal for our grade 7-8-9 mathematics pro- 
gram. Our student contact here is much more 
difficult to measure because junior high com- 
puter instruction tends to be much more infre- 
quent and diffuse. Still, we would estimate 
that the district impact here exceeds five hun- 
dred students and is rapidly growing. 

The results of the elementary program are 
difficult to assess. Only a few students have 
profited from their exposure to BASIC. This 
is understandable, however, considering the 
lack of computer training or knowledge posses- 
sed by most elementary teachers. It would ap- 
pear that the computer's future in the elemen- 
tary grades is as a learning device in the areas 
of gaming, simulation, drill and practice and 
tutorial instruction. The elementary program 
has been more successful where older students 
have been available to serve as cross age tu- 

With regard to hardware, we have found 
our microcomputers to be thoroughly dependable. 
Time between failure seems to be in excess of 
one year. Most of our machines are running on 
16K of RAM with 8 or 10K BASIC interpreters. 
Our system storage medium is audio cassette 
tape. This has proven to be quite satisfac- 
tory. Students store their programs on paper 
tape or in cassette cartridges. We are ex- 
perimenting with teletype vs. television mon- 
itor I/O and have reached no conclusions. We 
are finding our Cromemco Z2-D machine to be 
the most versatile, by far, of all our micro- 
computers. Students have their own mini-disk 
for program storage. Cromemco software (BASIC, 
FORTRAN, Z-80 ASSEMBLER, DOS) is superior to 

anything we have seen. Also the Cromemco 

machine can expand to eight BASIC users at a 
very modest cost for each additional user. 
Given the exceedingly low failure rate after 
installation, we expect to pursue this Cro- 
memco timesharing approach in our high schools . 
(The argument against timesharing is that 
when the machine dies, so do all the termi- 
nals.) We have also had a good experience 
with our one and only PET. It performs as ad- 
vertised and seems to be just as reliable as 
our other machines. Because of its attrac- 
tive cost ($595 for 4K user RAM and $795 for 
8K) we expect that we will purchase many more 
PETs over the next year or two. Since they 
are less versatile machines, they will prob- 
ably find their greatest use in junior high 
where computer literacy is the primary goal. 

There have been some problems, of course. 
We have found that a system should be thor- 
oughly burned in and tested before installa- 
tion. With microcomputers, the first few 
weeks of life are the hardest. We are also 
very much aware of the lack of good software 
interchange • BnSxu uas yet to uc standaruizeu 
and, just as important, there is no standard 
medium of exchange. 

In conclusion, we would like to emphasize 
that what we are doing is very "plain Jane". 
We are not engaged in anything at all exotic. 
We are mostly teaching BASIC. The context 
of our programming is largely mathematics 
because mathematics teachers have preempted 
the field. We like BASIC because it is inter- 
active and because it is so well suited to 
the range of abilities with which we have to 
deal. We know about CAI, but are not em- 
phasizing its use at this time because of the 
almost total absence of microcomputer adapt- 
able software and the high cost of preparing 
our own CAI software. More importantly, we 
are following the approach I have outlined 
above so that we can quickly adapt to the 
new demands being placed on the mathematics 
curriculum. Perhaps of most importance is 
our intense feeling that programming and 
computer science is a new curriculum area, 
related to, but somewhat separate from 
mathematics, science and electronics. The 
digital electronics has birthed a new sec- 
ondary school subject and we want to be part 
of its upbringing. 



BOX 1579. PALO ALTO CA 94302 


William J. Wagner 
Mountain View High School 
Mountain View, California 

Intr oductio n 

The title of this talk is meant to be 
ambiguous, for I will argue that two 
audiences are being expanded: the introduction 
of micro computers into the high school will 
allow computer education to be offered to more 
students, and further, educators like me 
represent a very different -set of users of the 
formerly hobbyist-oriented micro computer. So 
if you think my feet are shaking up here, it 
is merely because I am playing the role of the 
tip of the iceberg. 

Computer programs in schools are not 
new. What is new is that micro computers open 
the possibility of bringing computer education 
to many more students within the strictures of 
modern school budgets. 

This broadened market represents a 
real challenge for computer educators. In 
every school there exist some students (about 
1%, I think) who will, if you let then, spend 
all their waking hours at the terminal, making 
the machine jump through hoops, and 
occasionally emitting those familiar gales of 
laughter. They "are great people (they aire us, 
right?). Every teacher needs these kids 
around - they keep the adrenalin circulating. 
We are proud of their programs and like to 
take credit for their progress. Indeed, we 
should take some credit, but not for teaching 
them much. Our contributions are in 
suggesting interesting new kinds of problems 
and in getting the equipment in their hands 
in the first place. 

Important as it is to give these 
future professionals and/or hackers their 
first boost, our programs should not be judged 
or justified according to the progress of 
these superstars. Schools are full of lots of 
different people, and it is this other 99% 
which I intend to discuss here, and to which 
the program at Mountain View High School is 
directed. Also, although I am by no means a 
computer hobbyist and have only limited 
knowledge of the micro computer field, I will 

Home Address: 127 O'Connor St. 
CA 94025 

Menlo Park, 

present my views on the special requirements 
of high school computer facilities, and why 
we selected micros. 

Throughout this talk I will also ask 
for your help. Computer educators need better 
ways to find each other and share their ideas, 
plans, successes, and even failures, I hope 
this will be the beginning of a useful 
exchange among us . 

The Students 

Without risking more guesses about 
percent composition, here are five categories 
of students, for the purpose of discussion: 

A . The computer hotshots. These have been 
discussed above, but I should add that they 
are not necessarily the best students in other 
classes, and may even have trouble getting 
computer assignments in on time. We can 
really help them by focusing their energy and 
enthusiasm toward activities which will 
continue their growth. 

B .. The, goad., s Ludents_.-in__ma.th_. a nd science.., __. . 

These are our natural audience, but sometimes 
it seems surprising that they do not flock to 
the computer. For one thing, they are very 
busy with course work, and school courses are 
not often set up to reveal how the computer 
might be of use. And the tightly packed 
college prep schedule has little room for 

C. Other good students. This may be the 
largest group in any school. They also are 
busy with various school activities, but 
unlike group B, computers are not on their 
list of things to pursue some day. Further, 
they often show a remarkable disinterest in 
what we like best - computer games and 
programs which crunch numbers. 

D. Kids with lots of interests, but not 
particularly top students. Some of these 
might be interested in our hardware, but not 
so much in how to manipulate it. Others find 
programming to be their first interesting 



BOX 1579. PALO ALTO CA 94302 

"academic" class. Often their math level is 
so low, however, that the number of problems 
that can be suggested is limited. 
L. Unmotivated and unsucces sfu l students . 
This is another large group. They find their 
way into computer classes because their 
schedule is not full and because someone, 
usually not the student himself, thinks that 
something flashy and new may produce a change. 
It turns out to be true in some instances. 

I will mention in passing another 
group which has remained an enigma to me - 
young women. At my school they make up at 
least half of the enrollment in advanced math 
classes, but seem impossible to attract to 
the computer program. Those in the 
programming classes are every bit as good as 
the boys , but only one or two out of the 
twenty or so students who spend extra time 
programming or on games is female. I look 
forward to discussing this situation with 
persons interested in sex differences in math 
and science. 

The Importance o f Expanding Our Audience 

I think that we designers of computer 
curricula have tended to think in terms of 
groups A and B only. It is very difficult to 
find a programming book that does not assume 
knowledge and interest in higher math. Also, 
we have been content with the few students 
from those groups who flocked around us 
because we couldn't have served many more 
with our limited facilities anyway. 

It is important for us to reach out to 
this large majority of students who do not 
naturally find their way to the computer room. 
It is obvious that computers are actually a 
part of their lives whether or not they like 
or understand them. However, few 
opportunities will exist for them to interact 
actively with computers outside school. 

Furthermore, these students tend to be 
alienated from the computer activities at 
school - they think it is for smart people 
only, and they think that demented laughter 
and secret language means you have to be a 
little crazy too. (k colleague at a nearby 
school who brought her Algebra I class to the 
computer room for a week of introductory BASIC 
told me a wonderfully poignant story that fits 
here. As a 10th grade girl sat down at the 
terminal for the first time, she looked around 
nervously and inquired, "Can anyone see me in 

How can we reach these individuals who 
do not automatically think of computers as 
their thing? It is a difficult teaching and 
strategic problem, but we must continue to 

look for ways, for several reasons. First, 
this could be their last chance in a 
relatively non-threatening environment to have 
hands on experience with a computer. Once they 
leave high school their education and 
experiences begin to narrow toward a 
vocational or academic or lifestyle choice 
which will preclude taking a flyer into an 
apparently unrelated field like programming. 
Their contacts with computers will be passive, 
inspiring awe, fear, or anger. 

Also, the opportunities that do exist 
out there are more restricting. The college 
computer classes I know about are not intended 
for the liberal arts person, but for the 
future professional or technician. Thus we 
must think of some of our teaching as part of 
a general education. And for those students 
too busy, nervous, or skeptical to commit 
themselves to a full course, we must find ways 
to bring computers into the classes they do 

Another reason to teach these students 
about computers is that we can offer valuable 
experience in problem solving and thinking, as 
well as a new vocational direction. This area 
matches so many of the goals expressed by 
various departments of my school and of the 
school itself, that I am sure that if the 
slate were wiped clean and a new public school 
designed, courses about programming and 
computers would merit permanent status in the 

In this Utopian school computer 
instruction of some kind would be a natural 
part of most students' coursework: either a 
regular course like Geometry, the first step 
onto the college bound math track, or the 
equivalent of an elective in English or Social 
Studies, or a vocational offering like 
Woodshop or Auto Mechanics. I have found that 
experience with computers and programming can 
help teach math more effectively, can help 
students problem solve and organize their 
thoughts, and can of course lead to jobs at 
various levels of academic preparation. 

Revolutionary school reform is not 
imminent, so it is the responsibility of us 
educators to work for the gradual increase in 
use of computers in schools so that these 
benefits can be made available to more 
students and apparent to more decision-makers. 

Finally, I think that almost anyone in 
high school who is willing can be taught 
programming in BASIC at some level, and we owe 
it to them to provide this opportunity to be 
in control of the computer for once. The 
problem is pedagogical rather than one of 
prerequisites. What we need is a set of 
activities for various levels of student 
ability and interests. 



BOX 1579, PALO ALTO CA 94302 

It would be glib for me to say that 
this is an easy task, but I am busy working on 
the problem, and believe it can be solved 

The Cu rricular Offe rings 

At Mountain View High School we have 
tried to broaden the audience of our computer 
offerings by using five different approaches. 
They are a two week course in BASIC offered as 
a unit in certain math classes, simulations 
and games which have applications in various 
subjects offered in the school, easy access to 
the computer at designated times for game 
playing, classes in programming which offer 
experiences at three distinct levels of math 
ability and interest in programming, and a 
well attended adult education class in BASIC 
at night. 

First let's talk about what we do not 
do. There is no CAI per se because I have not 
figured out how to accomplish this in a class 
situation with our present facilities, and I 
await the software which can provide the 
organization, record-keeping, and pacing which 
are required in order for CAI to actually be 
a help to a teacher. Also, I await the teacher 
who wants to give it a try. Another thing we 
do not do is any kind of instruction about 
hardware or lower level languages. This 
is simply because I know almost nothing about 
these things. I am a teacher, a sometime 
programmer, and somewhat of a mathematician, 
but I have had a nervous awe of electronics 
ever since my younger brother became a ham 
operator in the sixth grade and began speaking 
in tongues. I hope that others will know some 
ways to approach these omissions, or will 
suggest other activities we could usefully 

The Two Week Course. This has been the 
backbone of our program. It has brought 
programming to a large number of students and 
it has been effective in building enrollment 
in our programming classes. Furthermore, it 
was a major factor in gaining permanent status 
for the computer program. 

This course is based upon two 
premises: (1) A remarkable amount of 
programming in BASIC can be taught in two 
weeks, and (2) if programming is introduced 
as a non-optional part of a course, then some 
of the barriers against participating are 
transcended (for example, none of the regulars 
are around to make people feel dumb, and most 
students tend to try harder when something is 
expected of them, compared to a situation in 
which the activity is optional). 

The two week course was first offered 

in 9 math classes during the Spring of 1977. 
The classes were Geometry, Algebra II, 
Trigonometry, and Calculus. Modified versions 
were given in three other lower level classes. 
I taught each course with the assistance of 
the regular teacher. I am confident that the 
sane proportion of students who pass these 
classes came to understand the essentials of 
the following statements: INPUT, LET, GOTO, 
PRINT, and IF.. THEN, how to work with strings, 
and how to use the system commands like GET, 

At the minimum each student wrote and 
ran 6 programs, some related to their 
coursework, one using strings to produce a 
conversation with the person at the terminal, 
and some of general interest like computing a 
batting average, or a grade point average. 
Let me emphasize that this was the minimum. 
In each class there also appeared advanced 
programs such as one similar to NUMBER, change 
calculators, prime generators, and, my 
favorite, a program which guesses the person's 
secret number. Each of the writers of these 
particular programs had never programmed 
before, and accomplished these results 
within two weeks. 

The facilities during this first go at 
the two week course were four teletypes 
time-shared to an HP 2000F at the Santa Clara 
County Office of Education. Everything was 
rented, because this was to be a pilot project 
to prove the value of the program and to help 
us better evaluate what facilities would be 
needed. I will discuss facilities later in 
more detail, but let me add here that four 
stations seemed sufficient, although one of 
the few complaints from the students involved"'"' 
the need for more terminal time. 

Games and S imulations . HANGMAN in 
French and Spanish (and also in Tagalog!), 
LUNAR in Physics, CIVIL and ELECT in History 
classes, story programs in Creative Writing, 
NEWTON (getting across the stream against the 
wind) in Trigonometry, demonstrating limits 
in Calculus. These are all activities which 
we have carried out with classes. We reach a 
lot of students this way, and a lot of 
teachers. There are more things we can do, 
and improvements we can make with these. We 
have not, for example, offered a coordinated 
unit using, say, the Huntington teaching 
materials. If others have, I would appreciate 
hearing from them. 

The problem of interacting with other 
teachers and their curricula is an extremely 
delicate one. We computer teachers must tred 
carefully - our programs and equipment must be 
easy to use (or we supply student assistants), 
and our offerings should be more than just 

iiEitr LiixeLS • j-ne eiaitc tidSo SitOUxu. uc awxt; 



BOX 1579, PALO ALTO CA 94302 

to get involved, and this may stretch our 
imaginations, given our limited facilities or 

Also, we must work to convince our 
colleagues that we can contribute to their 

1 net" l"UCt "• "1 Kooanco /l«->n'«- f f\r-rya.t~ t-^af fK fl ,r 

are probably not from either group A or B 
described above. I hope that other teachers 
trying the same things will get in touch with 
me, for this is an area of great importance to 
the increased success of our programs, and one 
in which sucessful strategies are very 

Game Playing. The computer facilities 
are open various times during the week for 
game playing by any student in the school. The 
HP system library is full of games, many with 
no educational value. We do not encourage 
these games, but they do attract kids. We try 
to show them the more interesting (from a 
teacher's point of view!) alternatives to 
FOOTBALL, BLACKJACK, etc, and there are many 
on that system. 

When we have changed over completely 
to our own computers there will be more 
control over the available games, but I still 
think that game playing will continue to be 
an important part of our program. There is a 
lot of thinking and problem solving required 
by the right selection of games, and I think 
such things are a useful addition to a 
school's extra-curricular offerings. Game 
playing leads some kids to inquire about 
programming, gives others ideas for original 
programs, and at the minimum lets a kid 
interact with a computer in a non-passive way. 

Programming Cla sses. There are now 
three sections of a one semester course in 
which the primary activity is teaching BASIC. 
In the course description I made Algebra 
"recommended but not required", which is 
consistent with my desire to reach more 
students and with by belief that programming 
can be taught to a wider range of people than 
is usually thought possible. Students from 
each of the groups A through E described 
previously enrolled in the class, and it was 
immediately apparent that the course could not 
follow a traditional format of lectures and 

The course now operates on three 
different levels. Some students just work on 
programs which they and I agree would be 
productive, and ignore the regular schedule of 
assignments. This schedule involves 
completing four chapters of our text, but 
almost every book or program assignment is 
divided into two levels of difficulty 
(students choose for themselves which level to 
tackle on each assignment). 

I would be glad to discuss the course 

in more detail with anyone interested, but let 
me close by highly recommending our text - 
Computer Progamming in the Basic Language, by 
Neal Golden (Harcourt Brace Jovanovich) - it 
is cheap and full of good assignments at 
almost every level * It does, however, assume 
a minimum of Algebra I, which has caused me to 
generate a lot of supplementary problems. 

Adult Education. A night class in BASIC 
using our facilities is being offered each 
semester. Right now the format is 6 weekly 
meetings of three hours each. I consider this 
an important part of our program, with 
potential for exciting growth. If you think 
about a public school as a community resource, 
many uses come to mind for a public computer 
facility downtown. We are interested to 
talking to others about these possibilities. 

Getting the Program Funded 

Of primary importance in the initial 
stages of our effort to establish a computer 
curriculum at Mountain View High School was 
the involvement of several dedicated parents, 
including one Board member. These parents 
insisted that the school should have a program 
comparable to the other schools in the 
District, and did not let the matter rest 
until it was a reality. Almost an entire 
school year was required to produce a 
proposal, and after another six months final 
approval was obtained to begin. 

In the spring of 1977 we began with 
four rented teletypes and offered the two week 
course described above. Its success produced 
wide spread interest in the computer 
throughout the student body, and led to a 
decision by the Administration and Board to 
make the program permanent. Other computer 
programs that I am familiar with in schools 
gradually grew over the years - a teletype 
here, renting or scrounging time there, with 
the continual struggle for marginal improve- 
ment each year. At Mountain View High School 
we now have adequate permanent facilities in 
large part because of sustained parent support 
in the early stages, and because we were able 
to involve a lot of students in a short period 
of time during the early part of the project. 

Computer Facilities 

Since October, 1977 we have been using 
two time-shared teletypes and two purchased 
Processor Technology "Sol" micro computers 
with North Star disk operating system and 
software. Each of the. units has 24K of usable 
memory, which means that BASIC programs may 
reach about 10K in size. We use an Okidata 



BOX 1579. PALO ALTO CA 94302 

printer which may be accessed by each of the 
computers through a centrally located switch. 

We have been very pleased with the 
Sol-North Star combination, and with the 
delivery and support provided by the Byte Shop 
of Palo Alto, As I write we are selecting two 
more units, one of which will be a Horizon 
computer from North Star with a dual disk 

The primary reason for choosing micros 
was cost, for the total purchase price of the 
four systems will be roughly equal to two 
years of rental of four time-shared teletypes, 
I was attracted to the DEC system which 
time-shares four ports, because of DEC's 
reputation and all the software which comes 
with it. But for their price we could 
purchase six of the Sol systems, and also 
DEC's maintenance costs over $2000 per year. 

Having become convinced that micro 
computers had ceased to be strictly a hobbyist 
domain, I made the plunge and have been 
delighted. Like the Holiday Inn commercial, 
the best surprise is no surprise, and for me 
the best system is one you don't have to think 
about. We load BASIC in the morning and can 
continue through the day with students 
programming, loading, printing, and saving 
programs. Often at lunch time, or as a reward 
for my General Math classes, however, the 
flashy Sol games are brought out. 

Computers in schools must be very easy 
to use. Students should be able to 
concentrate on the programming and ignore the 
machine itself. Teachers should not need 
special skills or knowledge if the use of 
computers is to spread to. more .schools. -and to 
more departments within the school. 

It is fine if one staff member can 
fathom the mysteries of machine language 
or hardware, but the overall program should 
not depend for its operation on this 
individual or his/her knowledge. Uncommitted 
teachers will not want to touch a system which 
seems inaccessible or arcane. 

In addition to satisfiying these 
unusual requirements for a computer whose 
tradition is with the hobbyist, our micros 
offer nice extras. For those students who 
wish to pursue assembly language, this is 
possible. The four systems are totally 
portable, and can each be dispatched to 
separate classrooms (or to different homes 
during vacations!). The FILL and EXAM 
commands (PEEK and POKE in some languages) 
bring a fascinating extra dimension to BASIC. 
And recently a student discovered how to 
allow data entry without the carriage 
return ... and so it goes. 

Our micros are symbolic of our entire 
program. Their cost, versatility, and ease of 
operation make it possible to bring computer 
education to many individuals who never would 
have had the experience, and yet they still 
can provide challenge and excitement for the 
brightest student. I am very pleased to have 
joined the ranks of the micro computer, and 
especially pleased to have been able to join 
on my own terms rather than yours. There are 
a lot of teachers and students like me out 
there, and I predict that next year this place 
will be too small to hold them all. 



BOX 1579. PALO ALTO CA 94302 


3304 Pico Boulevard 

Santa Monica, California 90405 

(213) 450-2060 

Formerly Director of Computer Activities 


2696 Valewood Ave 

Carlsbad, California 92008 

ffT tub itflRNlwr, f-flftM, 

We have been playing with our computer for 
year and a half now. All the students in our 
school use it in one form or another, as do 
the teachers. Less I mislead you, we have 
3-5 teachers and 20-40 students in a 'alterna- 
tive education' environment. Instruction is 
individualized, we try to maintain an 8 to 1 
student-teacher ratio. 

Needless to say, there was a great deal 
of resistance to the introduction of this new 
A/V technology to the school. I would get 
comments like "They scare me.", "I don't like 
computers.", and this was from the teachers I 
Distribution of a few biorhythm charts, compu- 
ter generated poetry, and Weizenbaum's ELIZA 
helped dispel any qualms about the system 
replacing teachers or taking over the school. 

Before I brought the computer into the 
school, I made arrangements with the teacher 
of a local community Junior College (Palomar 
College, San Marcos, California), Mr. Mike 
Michaelson to bring my students to the computer 
center a few hours a day for a computer program- 
ming class. Mr. Michaelson allowed access to 
the Univac 70/7 system and a classroom, much 
to the chagrin of the Computer Science Depart- 

My plan was to introduce the students to 
some computer games and gradually sneak in some 
educational courseware. Unfortunately, the 
college had no computer games, and frowned upon 

their valuable computer being used for such 
frivolousness. So, back to the drawing board 
While I was busy drawing up my new plan of 
attack - designing and programming the 
PILOT Interpreter/Editor System - unbeknownst 
to me, my students were busily typing in the 
code from 101 Basic Computer Games , DEC and 
What To Do After You Hit Return , People's 
Computer Company. They were doing this in 
the afternoon on their way home from school. 
Needless to say, this did not endear me to 
the Computer Science Department. 

However, Community College policy is to 
serve the community. When our plight was 
brought to the attention of the administra- 
tors of the college via the parents of the 
students our fairy godmother appeared in the 
form of Dean Cootz, Dean of Science and 
Technology and Tom Dolan, Head of the Computei 
Center. These gentlemen, through the 
Mathematics Department, offered us unlimited 
computer time for the duration of the semestei 
in order that my students might meet their 
high school requirements. We were required 
to allow the Business Administration students 
priority on the system. 

Remember the plan? My daddy always said 
"planning is essential, but plans are no damn 
good". Much to my surprise I discovered I 
had five computer nuts, now programmers, and 
their friends - potential computer nuts. The 
friends were playing the games that the 



BOX 1579. PALO ALTO CA 94302 

programmers had implemented. The games that the 
audience had chosen were STARTREK (now a classic) , 
This last game was of their own devising: take 
a game and reprogram it so that you have better 
odds, (there may be a lesson there). 

My daddy also told me "If you can't beat' em 
join 'em". It wasn't too far from our schools 
educational philosophy to let the students play 
games (after all, the teachers play some less 
constructive ones) , so I incorporated the games 
into my educational strategy. By this time PI/ES 
was completed so we started translating their 
selected games from BASIC into PILOT. The games 
that were not translatable, we used in some math 
context. The translation allowed the students to 
explore the fundamentals of computer languages, 
freeing them from a 'dialect' dependence. They 
learned personally how the implementation of a 
particular syntax forces a programming strategy. 

At this point, EDUTECH Project had agreed to 
provide to the San Diego County Department of 
Education one copy of the PILOT Interpretor/Editor 
System in return for the loan of a Teletype and 
access to their Burroughs 6700 system for a 

We still have access to the system and are 
at the time of this writing, negotiating for an 
extension in exchange for courseware. My thanks 
to Dr. Jane Gawronski, Mr. Bob Doolittle, and 
Mr. Bill Cue, not forgetting the glib operator 
with the sagacious one-liners that appear on my 
terminal at 3:00 in the morning. 

The final educational strategy was: 

1) Introduction via playing games, 
taught some basic skills: 


a) turning the system on & off, b) keyboard 
familiarity, c) loading and executing a 

2) Elements of Computer Programming. Using the 
PILOT language I introduce students to the 
concepts Input and Output. I give them the 
'T:', and 'A:' statements and have them 
write a simple program. 

3) The Black Box. Between Input and Output 
a Black Box that performs mysterious 

a) Move a string from Input to Output 

b) Remember something typed in. 

4) Decisions. The computer may compare an 
answer with an expected answer, and make 

a decision based upon this comparison. 
'M:', 'Y:' and »N:' instructions. 
(Match, Type if Yes, Type if No). 

Some of our students lose interest follow- 
ing number 4, above. They have the option 
of leaving the course at this point with 
a functional degree of computer literacy. 

5) The students are now ready for more 
esoteric concepts such as 'J:' - Jump, 
'U:' - Use (call), 'E:' - End (return), 
'C:' - Compute. 

6) The students are now writing and debugging 
their own programs and translating from 
BASIC to PILOT. The translation teaches 
them some programming tricks that are used 
by more experienced programmers, and how 
language capabilities differ. 

7) New tasks are introduced that are not 
possible with PILOT (such as array 
manipulation, File I/O, Function refer- 
ences) that occur in BASIC games. The 
students then dig into BASIC in ernest. 
In the same way, FORTRAN is introduced. 
Attempts to debug BASIC programs intro- 
duces the virtue of Structured Program- 
ming and Documentation. Our students 
discover programming on their own, with«c 
little guidance from the staff. 

When the other teachers witnessed the 
success I was having with the computer, their 
interest was aroused. The programming students 
introduced the teachers to the games that they 
had. written, and even taught_.thera the rudiments 
of programming. The computer terminal has now 
become the focus of school activities in the 
area of academics. 

One teacher wrote an English lesson pro- 
gram using the MADLIBS program in PILOT. 
Students interactively supply a list of adjec- 
tives, nouns and verbs of the proper tense 
and the program uses Mr. Basener's stony 'mask' 
to write a story. A student then wrote his 
own program to tell his own story. 

Mr. Tim Dawson, our English teacher, uses 
the Stanford Writing Programs by Ellen Nold an< 
Sally Cannom (as appear in People ' s Computer 
Co.) in his English classes. 

Mr. Peter Brown the Director of the 
Learning Farm, uses some PILOT programs 
acquired from Maria Montessori of the Golden 
Gate school in San Francisco for the younger 
children (thanks to Ursula Thrush) . 



BOX 1579. PALO ALTO CA 94302 

We have developed a few PILOT arithmetic 
and English Grammar programs. Our arithmetic 
programs get progressively more difficult and 
display the student's percent correct (grade) 
and average response time (time to answer a 
question) at each level change. The levels 
start with horizontal addition of two random- 
ly selected numbers less than 5 (5 + 4 = ?) 
and proceed in individualized increments to 
multiplication of two 3 digit numbers and 
columnar addition of three 3 digit numbers. 

The gimmick here is the response-time. As 
the students master the material (grade = 100%), 
they begin competing among themselves for quick- 
er response times. These routines provide self- 
motivated drill and practice. 

The grammar programs are also very simple 
so that there is a lot of positive action on the 
part of the student. One series of programs 
requires the student to identify a particular 
gr amma tical element in a given sentence, such 
as the verb. A second series requires the 
student to identify the part of speech of an 
underlined word in a sentence. A third requests 
a word of a particular part of speech, and then 
weaves the words into a story. In a fourth series, 
the student constructs a sentence from a list of 
words provided by the program. The program then 
checks the sentence grammar. These routines give 
the child experience manipulating symbols. Words 
become things - tools - to be used. 

In order to make this material available to 
users beyond the Learning Farm we are working 
on documentation, the hardest but most necessary 
part of the system. This paper is perhaps an 
element of that task. 

If you are interested in writing course- 
ware or lessonware, contact the PILOT Information 
Exchange at Box 354, Palo Alto, California, or 
the EDUTECH Project, Box 1023, Encinitas, 
California 92024, and you will receive some 
author's guidelines. 

Let me leave you with this advice - write 
the documentation first. 



BOX 1579. PALO ALTO CA 94302 


William P. Fornaciari, Jr. 
Math Dept., Polytechnic School, Pasadena CA 



There is, in any discipline, a fine line 
dividing serious work and play. So it is with 
computing in schools, and this problem (of 
defining serious work) is magnified by the 
large role that game playing and game 
development assumes in curricula with 
computers. It is inherent in interactive 
computing that CRT's tend to become 
recreational devices -- a visit to a 
university computer where students have 
virtually unlimited amounts of computer time 
on interactive systems (Caltech's PDP-10 or 
Dartmouth's facility, to name only two) will 
verify this. Games appear to be the sole 
activity on terminals (at least upon a general 
inspection) with little "serious programming" 
done at all (serious programming belongs 
exclusively to the batch processor, where 
there is usually a cash ante). With 
microcomputers being virtually universally 
interactive and video-based and the economical 
approach for use in secondary schools, the 
impulse to play games can be addicting beyond 
all reasonable value. After a student starts 
the nth run or version of Star Trek, how can 
you hold the line? 


Perhaps we might ask how many Klingons 
must be destroyed before a student ceases 
learning and is merely passing time, avoiding 
his or her academic assignments? The same 
question may be asked of an educational game, 
say the Huntington Project's "Pest 
Management" [1] : how many flies must the 
student wipe out before the game ceases to be 
educational and becomes solely recreational? 
For this program "as soon as s/he kills the 
flies cheaply" or "when s/he passes biology" 
might do; in general, "how much is enough?" 
might be answered by noting when 1) the 
student starts to modify the game; or 2) the 
student starts to modify the game, by 
resetting variables, to cheat; or 3) the 
student becomes bored with the game and either 
asks "What else have you got?" or gives up 
computer games and wanders off toward other 

The game is a means by which we attract 
people to computers — and perhaps hand out a 
little ethical education. In any of the 
situations of extreme behavior with respect to 
game playing or, sometimes, game development 
(at least of those games which we might 


consider with little educational value or 
simply variations of the same theme) , the 
educator must reevaluate the role games play. 
Surely there must be other activities. I witto 
to present some experiences and observations 
as a preamble to a short forum on alternatives 
to games which will follow this talk at the 
Faire. Last year, Liza Loop f2] observed: 
"Don't introduce Startrek or gambling games in 
class. You'll never get the kids' attention 
back." Some of my observations are as 
profound, but designed for those who did not 
heed Ms. Loop's advice. They generally apply 
to older students (most students where I teach 
math and supervise computer activities have 
virtually unlimited time fro grade seven 
through twelve). For standardization, 
computer statements are in BASIC. 

Getting Started: Problems 

Without categorically rejecting games and 
their development in a curriculum, there must 
be some other activities the teacher can 
suggest as an alternative to trekking. Perhaps 
we need only look at the batch processor for 
the answer. Most will agree that time spent 
programming a 370 (or whatever) is time 
well-spent (never mind that the computer money 
might not be better spent in other areas of 
research or __ other approaches) ._We batch 
process because we have a problem to solve . 
Why can't this use be fostered in high 
schools? It is important to preserve the 
integrity of of the computer education 
department as a competent entity of problem 
solvers and not game players. Then, and only 
then, will we be asked by our colleagues to 
solve problems, and maybe develop a few. It 
is, of course, not as much fun — at least at 
first. And, upon running short of problems, 
other approaches can be suggested. If work 
seems justifiable, then the time spent on 
games seem to be a little more justifiable — 
some time off for a brief period of 
recreation. I'm not saying no one should play 
games or write new ones for computers; I've 
indulged quite a bit in this pastime and it's 
great. But other things can be achieved with 
as much pleasure, and it is important to 
develop a working spirit. 

Game development reveals a lot : one of 
the most illuminating and frustrating 
afternoons was spent with a particularly 
bright and very imaginative ninth-grader whose 
descriptive abilities had not quite matured. 

206 BOX 1 579. PALO ALTO CA 94302 

He was interested in programming a hide-n-seek 
game of the twenty-second century, where 
players took turns peeking from behind a 
barrier, semi-permeable to laser (or was it 
phaser?) pulses; you got so many chances to 
move or lase, or whatever, until you blew your 
opponent (he hoped would soon be the computer) 
off the face of the screen. "Whew!" I 
declared, "that's some game, but let's play it 
together without the lasers, please." With 
difficulty avoiding reference to the 
futuristic ordnance, this esger fellow (who 
happens to be a good chess player, good at 
math and not a computer novice) to agreed to 
show me his game with a chessboard, a couple 
of pawns, an invisible barrier which we both 
agreed to respect, and notepaper to record the 
results and rules. As it turned out, the game 
was totally undefined, von Neumann would have 
been both pleased (it's clearly a zero-sum 
game) and disappointed (it has no rules) . I 
keep a paperback on game theory near the 
computers, and most are, needless to say, 
dissappointed at its definition. Corollary to 
this, it's a good idea to have a game 
development system: checkerboard, dice, cards, 
a copy of Hoyle, etc. near the computers. 

The above illustrates one of the 
fundamentals that holds computing, and the 
methodology of clearly defined and logical 
procedures we subscribe accompanies it, 
together: if a person cannot verbally describe 
an operation, s/he surely cannot write a 
program to perform it. Problem definition is 
a constant problem (but the effects of its use 
are strongly felt in computing and elsewhere) 
— when is requesting a problem definition 
asking too much? At one point, requiring 
students to request time with a clearly-stated 
paragraph was tried for a couple of days, but 
was not too warmly received; it was hoped this 
would weed out the monopolizers and those who 
were constantly writing trivial programs (10 
PRINT "HELLO"; :GOTO 10) as jokes. At best, 
you can pester students to do this, and to be 
self-consistent, this would require the person 
who might be curious enough to begin computing 
to request, in writing, to play MUGWUMP [3]. 
At worse, the whole operation might succeed, 
and the teacher becomes a miniature 
bureaucracy awarding time for successful 
proposals (this might be educational, but 
let's defer this lesson for as long as 
possible) . 

Working with the Students 

Let your students define with you the 
operating procedure for allocating computer 
time (I've tried limited hours per week by 
sign-up sheet and it's really hard to 
maintain) . Agree to practically anything, 
including unlimited time, but insist on each 
student's having at most three programs or 
projects (two is better) that s/he might work 



on. Then when the project becomes bogged 
down, insist on clearly-defined questions and 
problem definition before you help them out 
(you won't be able to do so any other way). 
The only escape from this, for the* 
undocumenting student, is help from others. 
Reserve the right to restrict this, and you, 
and your trained students, will straiqhten out 
the corner-cutters. 

For those that begin by typing in games 
from the various sources, suggest that they 
limit the associated story to skeletal lines 
and go for the computation statements (note: 
it is imperative that linenumbers be preserved 
in order to assist in proofreading and 
debugging. Also, remember to acknowledge to 
source, author and typist/translator in a 
comment; early adoption of this practice will 
insure later respect of authorship and 
documentation — comments may be minimized, 
but not removed, in later practice an 
original program should begin with comments of 
declarations, entry points, subroutines. The 
program will then seem to write itself.) 
Once a student has completed the guts and is 
satisfied that the game is meaningful, it is 
appropriate to append instructions, fancy I/O, 
etc. Maybe s/he will find the game was not 
all it was advertised to be; s/he will have 
saved some time (lots if s/he doesn't know how 
to type) or possibly have found something more 

On the other hand, a BASIC INPUT 
statement without a prompt message is 
worthless. When posed with a solitary "?" 
during the execution of a program, I am 
tempted to respond to this ERROR 477 [4] with 
"WHO, ME?" and await the usual "REDO FROM 
START" (unless someone was just named "WHO, 
ME?") . While the computer misses the message, 
the programmer seldom does. Error recovery is 
a very important habit to develop. It is not 
to much ask for idiot-proof programs, with 
software designed to work with humans. 
Programs are complete when some proof of 
correctness has been demonstrated; we must 
instill the importance of documentation in all 
students . 

Assembly Language 

Z Assembly language coding is surely one of 
the best activities for those who really 
understand BASIC to the point where they're 
almost bored or keyboard in short programs but 
fail to develop them. To be sure, programming 
m assembler requires time and dedication on 
the part of the teacher, and usually some 
classroom exposure to the instruction set. In 
some circumstances, this may be impossible to 
provide directly, but for those who a true 
prodigies, the manufacturer's monitor listing 
and the chip's manual may be all that is 
necessary, with a few video games from Dr. 
Dobbs' (here the means justify the ends) or 

BOX 1579. PALO ALTO CA 94302 

other source [5]. Programs, utilizing 
memory-mapped output via a VDM or VTI , are 
especially good , because the results can be 
quickly realized. Even if the program blows 
up, some video output has probably evolved, 
giving hope, necessary according to Joyce. [4] 

Start students assembling by having them 
copy programs and patch I/O to your own 
system; pretty soon they'll come up with their 
own ideas, or possibly assist in getting some 
major systems software implemented. Mostly, 
by the nature of assembler, they'll be 
diagnosing errors away from the computer , 
defining their problems more carefully. 

At present, I have a half-dozen students 
working on assembler programs. Excluding the 
student who speaks macros, two are working on 
programs in assembler because it's the only 
way to achieve the speed (ever try to write 
PONG in BASIC using POKE statements?) , one is 
interested in robotics, and several others see 
it as an opportunity to increase their machine 
time, though they don't realize it is actually 
freeing the machines, while puting greater, 
but tolerable, demands on my time. In the 
end, I think, it should pay off as T get six 
reliable programmers to implemet software, 
debug, document and maintain the hardware. 
It's a chance to work more closely with the 
students, and sometimes you wonder who's the 
teacher. They're full of ideas. 

The Unexpected from the Youngest 

In the summer of '77 (which will be best 
remembered for "Star Wars" and a demand for 
more spectacular space games) , I assisted Dave 
Kressen, a veteran, educator ., math, and computer 
science teacher of the junior high (and 
younger) levels, in a "summer school" offering 
of programming micros for students post fourth 
through eighth grade. Dave, with several 
years' experience using Cal tech's PDP-10 as 
the principal computer, had taken task to 
write his own BASIC primer, "The Mathematics 
of Programming in BASIC" [6] , the previous 
year, compiling a dozen arithmetic problems to 
be solved, each adding successively complex 
programming techniques (I generally refer my 
senior high students to this text and suggest 
they try the programs to see how counting 
methods are used, deciding the best conditions 
on IF statements with the multiple statement 
per line BASIC used by micros; Kressen's 
booklet uses ANSI BASIC and hence, is not code 
optimized). For the most part, the older 
students were interested in games, Star Trek, 
and cornering the market on source tapes which 
they might trade at a cash profit, which is a 
problem to be solved. The new fifth and sixth 
graders, having a first "hands-on" experience, 
were going through the mathematical exercises 
and, with only a slight discpplinary hand, 
tending to business. 


When the problem of generating primes by 
Eratosthenes' seive was assigned, quite an 
amazing transformation took place — the 
younger students were absolutely spellbound by 
primes (it should be noted they understood 
what a prime was) , and especially by the rate 
at which the gaps between primes grows. There 
was considerable speculation about when a gap 
of 100 might be found. Tt became necessary to 
dedicate one computer to generating primes 
overnight; this required a premature 
introduction to the one-dimensional array (to 
hold the primes between print-outs) and the 
two-dimensional array (to hold the lower of 
two prime and its associated contending gap) . 

The kid's ideas and questions on this 
particular of how to generate primes faster 
(Kressen's original program used trial 
division by all odds less than the square root 
of the odd integer in question) were 
unbelievable. Ten-year-olds suggested that it 
would be faster to try only known primes as 
trial divisors, and thus a bootstrap routine 
to generate the first primes to 1000 was 
written to produce a basis set. The problem, 
designed to introduce the GOSUB statement, 
captivated the class for over a week, started 
programs in prime-factorization, tests for 
perfect numbers and eager efforts on 
Goldbach's Conjecture (that all evens can be 
expressed as the sum of, at most, two primes). 
When asked about the apparent slow speed of 
the computers (a weekend to generate all 
primes between 200,000 and 900,000) most were 
able to identify the method of programming as 
the ultimate limitation and several suggested 
that the numbers could be represented as whole 
numbers rather than reals to speed things up. 

Motivation and Incentives 

Within most fields, incentives to produce 
a superior product are commonplace, arising 
out of competition rewarded by recognition . 
Practically every aspect of education offers 
incentives, competition and recognition, and 
so it should be within a computer curriculum. 
Certainly writing and debugging a particular 
program are milestones themselves, affording 
the student the satisfaction of mastering a 
demanding servant. Perhaps, also, the numeric 
result achieved will find its place among 
others in a completed piece of research or the 
cataloging of records or similar data 
processing effort. 

Building a library of good software 
requies that we provide these three 
ingredients to the student programmers. An 
immediate recognition of a minimum level of 
quality occurs when others ask for copies of 
programs, and for works of particular value, 
requests that they be made universally 
available (with appropriate documentation) 

BOX 1579, PALO ALTO CA 94302 

introduce new incentives to the student to 
improve the next effort. We must always 
challege his or her capabilities in new 

A particularly effective recognition is 
to ask one competent programmer to write a 
particular program. In this instance it is 
particularly important to approach the student 
as a professional. S/he may or may not expect 
clearly specified I/O routines, but after 
having received them, s/he will appreciate 
their value. Commission your better 
programmers with good problem definition, some 
suggested reference materials or algorithms, 
and be surprised by the results — it can even 
be done with novices, if you offer a little 
help. Two students wrote an effective class 
grading program as their third BASIC project 
in only several hours after some preliminary 
meetings about sorting methods, averages, 
medians, data allocation and trying two 
programs out of Kressen's book (they were able 
to translate mathematical algorithms into 
character algorithms after agreeing that "A" 
is no more equal to "B" than 1=2). 

The concept is no different than giving a 
programming assignment to a classroom of 
students and assigning grades; in this 
instance, however, you deal with one or two 
(that's the "gradelike" incentive) and follow 
up by using the commission work, saving it as 
an example of good, complete work which might 
be the basis for the next program 

A modification occurs when the school is 
part of a user's group. We are part of a 
group which has an excellent barter economy 
established. For each program accepted, we 
get box tops worth $15 in exchange programs 
(which usually sell for $3 to $5, which seems 
like a fortune to some kids.) It means we 
get recognition for contributing worthy 
programs, we can save some effort keying in 
programs we don't already have and we get to 
see what others schools are doing with similar 
systems. The biggest problem is a common 
medium of exchange. For users of identical 
equipment, there is little problem, but 
Grimes' [7] observation has been prophetic. 
Schools should be at the head of users groups 
— there are too many opportunities and 
incentives lost without them. The programs 
should not be limited to BASIC, as assembler 
and CAI programs need encouragement and 

Finally, end-of-year recognition is no 
more out-of-place in computing than in 
basketball or scholastic achievment. 
Certificates of Excellence should be awarded 
by a knowledgable faculty/professional 
committee and should be vertically 
distributed. The incentives, direct and 
indirect, with adequate recognition will 
produce good competition. The combination 
result in a productive atmosphere where the 
students feel they are truly working, insist 

on conditions suitable for productive work and 
become discriminating and efficient in their 
use of time and resources. They will also see 
the computer as a tool to solve real problems, 
because, in the atmosphere described, that is 
what they hve been doing throughout. 

A Summary of Suggestions 

The following might be useful in 
developing and maintaining a computing center 
where little question of the justification of 
its existence will ever arise. Some 
suggestions have been developed above, and 
others are offered without comment: 

1) Computer educators must encourage 
colleagues in other academic fields to assign 
problems which require computation to achieve 
a satisfactory solution. Numerical solutions 
to long algebraic problems, simultaneous 
equations, even atomic structure and 
solid-state problems become reasonable but 
challenging for high school level science and 
math, as examples; social sciences can use 
simulations and statistics which are really 
just games with a lot of data. 

2) Get and use an assembler. The 
problem solving techniques required to use 
assembly language are applicable in any field; 
furthermore, you'll be able to develop and 
implement a very extensive library of utility 

3) Start and maintain a library of 
programs, a use your students to get the job 
done. Insist on documentation before and 
after actual programming. Encourage 
competition and show recognition when earned. 
Participate in an exchange with other 

4) Try to promote the idea of data 
processing at the student level, i.e. maintain 
club rosters, newspaper advertising accounting 
on the educational computer. Have students 
write the programs. 

5) Maintain a workroom atmosphere 
relative to the student age level. There will 
still be some frivilous game playing one, but 
usually as a temporary diversion. 

If you must Trek , by all means insist on a few 
basic modifications to the probably most 
widely-found version by Lynn Cochrane [81 . 
The course bearing notation (l=east or right 
to 8.9) is found nowhere on (even 
astronomical) maps. Use the standard polar 
coordinate system where degrees is to the 
right; try accepting angles greater than 360 
and negative bearings (figure another way to 
abort the torpedo or engine command) . For 
advanced students, have them indulge in radian 
measure (in units of pi), and, memory 
permitting, replace Spock with an "on-board" 
computer which returns principal angle 
arc-tangents given rises and runs (remember 



BOX 1579. PALCKALTO CA 94302 

the division by zero) ; the students can then 
think to figure out the proper quadrant 

Also, I'm a little incredulous about a 
photon torpedo's (even one from the 
Enterprise) capability to destroy a star — 
disable this awsome feature with a hard fix. 
To do all of this, you will have to look at 
the program. "What to Do After You Hit Return 
[9] makes its biggest point in the value of 
game playing when it suggests the game is not 
so important, it is how it is play and how it 
is programmed. You'll have to figure out how 
Cochrane makes the stars go away before you 
can save them. So for all the problems Star 
Trek has created, here's the praise it 
deserves as being a model program of 
efficiency. A good starting assignment for 
your students: explain where Klingons come 
from! And if they waste those torps on stars, 
or spend afternoons stalking Darth Vader, 
insist they be quiet about it!! Others are 
working !! ! ! 

Conclusions : 

I have attempted to define several areas 
that can be considered productive activity and 
those which are recreational. To justify a 
computer activity as productive is very 
subjective, but a reasonable definition might 
be an activity requiring that an active mental 
process employed to achieve a defined goal — 
work, if you will. And if your goal is to 
destroy 23 Klingons in 30 years, to do so 
might require an active mental process, though 
usually not for long. If given challenging 
problems to solve, encouragement and a little 
incentive, students will opt to solve 
problems. And "problems" are relative, also. 
To a ten-year-old, his or her curiosity might 
be the biggest problem. To an older student, 
using a different language to represent the 
same data and procedures is truly a 

One person's play is another's work — 
how or where do you draw the line? It is an 
important question to ask yourself, and to 
reflect thought and development in your 
computer curriculum. 

is as vague as it appears here. 

5. There are too many references to list, but 
Marvin Wizenread's video games, in several 
issues of Dr. Dobbs Journal , are particularly 
good, as well as reasonably short. 

6. David P. Kressen, "Mathematics of 
Programming in BASIC." unpublished. 

7. Peter S. Grimes, "Classroom 
Microcomputers." Proceedings, p. 165. 

8. Lynn Cochrane, INTERFACE magazine, July, 

9. What to Do After You Hit Return People's 
Computer Co., Box E, Menlo Park CA 94025. 


1. Huntington High Scool Project 

2. Liza Loop, "Sharing Your Computer Hobby 
with the Kids." First Computer Faire 
Proceedings, p. 156. 

3. MUGWUMP is a number guessing game, see 
ref. 9. 

4. James Joyce, "Human Factors in Software 
Engineering." Proceedings, p. 56. ERROR 477 

BOX 1579. PALO ALTO CA 94302 


Richard Harms, Santa Ana College 
Santa Ana, CA 92706 


It is interesting that we 
should have in our language the 
idiom "A person learns a bit at a 
time". This paper describes an 
instructional approach based upon 
the micro computer equipped with 
tape cassette. The approach is 
structured around the premise that 
"A person learns a bit at a time". 
Small segments of any topic are 
presented to the student. The 
student works with the material at 
his own rate. Based on his responses, 
individual learning pathes are 
developed. Non-computer disciplines 
including accounting, journalism and 
psychology have been successfully 

The Rationale 

The problem of designing learn- 
ing packages for computer is not a 
new one. Work in CAI (Computer 
Aided Instruction) goes back over 
the past ten years. The cost per- 
formance characteristics have tradi- 
tionally been achieved by spreading 
an expensive resource (the computer) 
over many users and long periods of 
time. It is then generally possible 
to show favorable cost performance. 

With the microcomputer the 
primary cost constraint in CAI is 
eliminated, namely the expensive 
resource. We can then concentrate 
on what should be our primary object- 
ive of computer learning - effective 
learning. It is this objective with 
which we have attempted to work at 
Santa Ana College, and this is the 
nature of my presentation. 

Within the confines of the 
latest widely marketed microcomputer, 
a system has been developed for 
authoring and presenting CAI which 
requires very little computer ability 
of either instructor (author) or 
student. This system is an outgrowth 
of a successful system called GULP 
(General Utility Language Processor) 
which was implimented several years 

earlier on an educational mini- 
computer . 

As is true of other CAI pack- 
ages, two distinct processes are 
required. The first is the author's 
creation of the material and the 
second is the delivery of the mat- 
erial to the student. Both pro- 
cesses are discussed below. 

Creating the Lesson 

As a general level of soph- 
istication within the microcomputer 
state-of-the-art, learning programs 
can, at most, be tutorial in nature, 
including sufficient opportunity for 
student drill-and-practice. The 
microcomputer of today by its very 
intent cannot be used for extensive 
simulation and gaming learning 
models. To prepare the tutorial 
lessons, then, the instructor must 
identify salient points which are 
student important. Prior material 
assumed known or unknown must be 
clearly defined. The instructor 
prepares briefs of each background 
point. Physical constraints such 
as screen size are considered in 
writing the briefs. Briefs of 
student important points are also 
prepared. Meaningful interactive 
questions are developed for each 
brief. Consistent with the initial 
assumption that learning takes place 
"a bit at a time", each fragment is 
kept discrete. 

After the instructor has 
created the text and questions and 
answers for the student learning 
session, an interdependent relation- 
ship is established manually between 
the lesson fragments. This is, in 
fact, similar to the PERT process. 
From these dependencies a linear 
string of fragments must be produced, 
If ties exist, an arbitrary ordering 
must be made because the next step 
in the process is the transfer of 
all fragments to the tape cartridge 
device. The fragments are serially 
chained. Subsequent fragments are 



BOX 1579, PALO ALTO CA 94302 

response dependent. A correct stu- 
dent response may cause the program 
to route to one fragment, an incorrect 
student response may cause the program 
to route to a different fragment. 
This allows the most flexibility for 
developing materials geared toward 
individual differences. 

The approach is not without 
flaw. If it is necessary to deliver 
the same material twice (i.e. the 
student did not "get it" the first 
time) , the material must be stored 
twice on the tape. Good lesson 
design will keep textual material 
brief enough so that the material 
is still on the screen if the question 
immediately following the material 
is missed. Using this technique, 
the instructor can tell the student 
of the error and refer the student 
to the material still on the screen. 

The program for storing the 
learning fragments, prompts the 
instructor, accepts his input, and 
stores the lesson on the magnetic 
tape. All the instructor materials 
are stored as data on the tape. In 
delivering the data to the student, 
the structure of each lesson is 
such that the delivery program is 
identical for all lessons. 

terminals serve as a great help in 
providing interactive information. 
Not all students attend every class 
every day. We have found the stu- 
dents much more willing to stay with 
us to the end of the semester when 
they have a place where they may 
pick up what they have missed. We 
are very optimistic of the future. 
We look forward to larger micros, 
more capabilities, less expensive 
quiet printers and numerous tech- 
nological advances which will continue 
to make the microcomputer the most 
exciting educational innovation in 
ten years . 

Presenting the Lesson 

After the instructor has pre- 
pared the lesson and stored it as 
data, multiple copies of the tape 
are made.. Tape costs are low and 
the intent is to have available in 
the classroom a copy of any lesson 
for any student at any time. We 
are incoroporating the microcomputer 
into the classroom in much the same 
way that typing classes have trad- 
itionally provided a work station 
for each student. The class is 
conducted (in the traditional way) , 
using the lecture method. Students 
may concurrently follow similar 
presentations on the microcomputer. 
Alternately, if the classroom pre- 
sentation is inappropriate for the 
student (too easy, too hard, re- 
dundant, boring, etc.), the student 
is free to pursue learning at the 
terminal. Some time is allocated 
for learning only from the micro- 
computer. During this time the 
instructor is free to circulate 
about the class answering questions 
and helping on an individual basis. 
Finally, outside of class time, the 



BOX 1579, PALO ALTO CA 94302 


David M. Stone, Teacher, ESS D 
Box 932, Pacifica, CA 94044, 415-589-5900 

The current press to get back to 
basics of education (The 3 R's) comes 
at a time when I, an elementary teach- 
er, can get funded to assemble a com- 
puter for use in the classroom. With 
the help of students and other teach- 
ers, we have developed a program use- 
ing BASIC (Beginners Ail-Purpose Sym- 
bolic Instruction Code) , one of many 
computer languages. 

Three "Computer Program Opera- 
tors", who are students from my fifth 
grade class, set up the computer dai- 
ly for the seven classrooms of 2nd, 
3rd, 4th, and 5th graders who are us- 
ing it this year. My suggestion to 
name them three CPO • s seemed quite 
logical, I thought. My "Star Wars" 
conscious students named themselves 
C3PO's, politely, seeming to have 
agreed with me and correcting my 
misordered science fiction name all 
in the same breath. (It was days 
before I realized that it was not my 
suggestion to which they had agreed.) 
Their diplomacy and understanding has 
been important to all of us involved 
in the project. 

For those of you who are devel- 
oping computer projects or are warming 
up to the idea of bringing your local 
students some familiarity with this 
marvelous tool, I hope I will pass by 
closely enough to what you need to jog 
the right ideas to mind as it seems to 
have happened with the C3PO's. 

The Need 

First in importance, perhaps, in 
developing a computer project is moti- 
vation. When working with a class of 
thirty in fifth grade math, the mathe- 
matical ability of the students may 
range in an approximate bell shaped 
curve from second grade to high ninth 
grade with several of them grouped a- 
bout fifth grade in ability. Chances 
are no two students would have the 
same set of strengths and weaknesses. 

As a beginning teacher in fifth 
grade, I learned that the text book 
the students were using was excellent 
for developing concepts but short on 
developing computation ability. In 
the following years the only thing I 


added to the computational part of the 
program was a drill in the math facts. 
Class averages in the ensuing years 
came up substantially. The importance 
of the drill in rote memorization of 
the math facts came home to me one day 
when checking how one student was do- 
ing, who always took about twice as 
long as anyone else in completing his 
drill but almost always got a perfect 
paper. He was in the process of count- 
ing five times five on his fingers 
under his desk. It took him 20 minutes 
at first to do the 40 most difficult 
multiplication facts. During the year 
his time gradually came down on the 
weekly drills to one fourth of his orig- 
inal time at the beginning of the year. 
He came back recently and said that he 
had continued to do well in math in 
junior high. 

It took 20 to 30 minutes to grade 
and record those tests on the math fact* 
in multiplication and division. Not 
everyone had the tenacity to stick to 
the drill and attempt perfection as thai 
one student did. A fair number of 
students had reoccurring difficulty onl} 
with certain math facts while knowing 
others quite well. As you can see it 
all boils down to a situation that the 
computer can handle outstandingly well. 

Implementing the Use of Hardware 

Some students just need enough timt 
to finish in order to improve. Some 
just need to have more drill on certain 
facts. Others could use the time bette] 
by occassionally passing up the drill 
for developing other skills. The set- 
ting where each student works at his owi 
rate and level of difficulty is called 
individualized learning and is getting 
considerable attention in the field of 
education today. 

Enter, microcomputer nearing the 
end of its second year of being bumped 
over deep cracks in our sidewalks as 
it is rolled from room to room on its 
little AV cart. The children can see 
the computer through the clear plastic 
on the lower shelf which protects it 
from accidental injury. As the compute: 
rolls into the classroom the teacher 
continues to teach. The computer oper- 
ator plugs in the computer and the TV 

213 BOX 1 579, PALO ALTO CA 94302 , 

monitor, gets the disc from wherever 
the teacher keeps it in her classroom, 
inserts it in the floppy drive, and 
starts up the first program which es- 
tablishes the date for that day. As 
the monitor leaves, the first student 
in a predetermined order moves to log 
on the computer and the computer has 
chained from the date program to the 
math program. If the student is an 
intermediate grade student the com- 
puter will ask for his last name and 
an identification number. If the 
student is a primary grade student 
the computer has asked whether that 
student was there. If not the com- 
puter asks if the next student is 
there. If the next student types in 
uyii or "yes" the computer asks for 
his identification number. When the 
last student is finished the computer 
asks once more for each student who 
was absent the first time around. If 
time permits the computer also chains 
to an educational game which develops 
some concept such as "is greater than" 
(>) or "is less than" (<). 

At the end of the morning or af- 
ternoon the computer operator comes 
in to remove the computer and returns 
the disc to where he found it. Mean- 
while the teacher has continued to - 
teach uninterrupted. Even in the low- 
est grades a computer operator aide 
can be found to correct small problems 
like "input error" or call the upper 
grade operator if it's something that 
can't be corrected by the keyboard. 

One second grade substitute 
teacher I talked to one morning was 
very fearful of taking on the computer 
especially since a new logging on pro- 
cedure was to be used that morning. 
I explained the procedure to one of 
my computer monitors in a few seconds. 
He explained it to his assistant and 
was not called back the rest of the 
morning. At noon I asked the sub- 
stitute how it went. She was all 
smiles and relief. She said, "Great! 
It was simple! " 

At the end of the day or whenever 
she has time the teacher takes her 
disc down to the teletypewriter to get 
a printout that looks something like 
A-l. Some teachers just send the disc 
down for a print out. 

The Use 

The primary concerns of the pro- 
gram are that the students always 
achieve an adequate rate of success 
and that the administration of the 
program is acceptable to the indivi- 
dual teacher. To the low achiever 

success seems to be a strong motivator. 
To the high achiever success and an 
adequate challenge seem to be the key. 

At this writing it seems we are 
able to meet the needs of all students. 
This is done in two ways. First, 
through variables in the program that 
automatically move the student from one 
level of difficulty to the next which 
are chosen by the teacher before the be 
ginning of the year. Second, other var 
iables may be changed for an individual 
after he has shown a need for more time 
or a different kind of problem. For 
example, the teacher may change this 
variable by herself by loading and 
running a program which gives her con- 
trol over any variables which she may 
want to adjust for an individual studen 
or perhaps the whole class. 

You can see from the preceding re- 
port that the teacher is relieved of 
some time in correcting papers, making 
dittoes, and running them off (which 
also saves paper) by using a computer. 
The teacher's valuable time and ex- 
perience is moved away from clerical 
work to focus on evaluation and even 
encouragement. (At the end of the 
student's drill, one of perhaps several 
comments tailored to his or her person- 
ality, could be printed out.) Control 
of the students' educational welfare 
is, therefore, comfortably in the in- 
dividual teacher's control. 

If what you would like to do is 
somewhat like what I have just describ- 
ed, some of the following hardware 
thoughts may be important to you. 

Hardware Considerations" 

If funds are limited, get a com- 
puter which will be able to handle time 
sharing. When you get the money to 
expand to more terminals you can do it. 
Almost 200 students use our computer 
each week at my school. This gives 
them about two minutes each. (A fifth 
grader can do forty multiplication 
facts in that time.) Two terminals 
would give each student twice as much 

time. Three terminals would 

well, you can see my point. 

The importance of some sort of 
timesharing hardware comes home when 
you only have to sit down at one ter- 
minal to ask only one disc to give you 
a listing of the events of the day 
from your students. Remember, comput- 
ers are here to save you time and work. 

A timing routine or hardware de- 
vice is important. Students who don't 
get on during the day because someone 
else is very slow get upset and they 
upset the slow student. Lincoln Semi- 



BOX 1579. PALO ALTO CA 94302 

conductor of Sunnyvale, CA has a rea- 
sonably priced and, from my experience, 
reliable timing board. What I'd like 
to find though, is an affordable board 
that doesn't need to be told the month, 
day, hour, minute, and second when the 
computer is first turned on. Something 
powered by a small battery might do the 

The Future 

One hardware note for the future : 
keep an eye out for interactive video 
discs. Phillips and MCA are working 
independently on players for the ed- 
ucational and industrial (EIT) users 
(1). Conducting a discussion while 
using a computer controlled video 
display could be just a year or two 
away. Imagine controlling the ani- 
mation, as it happens, to illustrate 
your point. It may be the most en- 
gagingly happy experience you or your 
students have enjoyed in school. 


(1) Braun, Ludwig, Video Discs : 
Magic Lamps for Educators? , 
p. 14, People's Computer, Vol. 
6 No. 4, Jan. -Feb. 1978, 1263 
El Camino Real, Box E, Menlo 
Park, CA 9402 5. 



Teacher ?Stone 

Is this a parent 



January 14 









1 Randall S. 

2 Dave S. 

3 Don D. 






1: 29 
2: 1 
1: 20 





1: 10 
1: 34 
1: 20 



COR = Number of Correct Problems 
TOT = Total Problems Attempted 
% = Per Cent Correct 
TIME = Time on Drill Exercise 
SGN = Sign of operation to be done 
AV% = Average percentage of all attempts 
AVT as Average time of all attempts 
RND = Round or number of turns at the 




Kelvin L. Zeddies 

1854 Pacific Beach Drive 

San Diego, CA 92109 


It is well known within the computer 
science community that computers are 
trans-disciplinary. They provide, 
in many ways, a unifying influence 
for educational experiences. Now 
that the shock of personal computer 
system availability has subsided, 
we must, develop and implement 
programs in the schools that provide 
students with experiences enabling 
them to use personal computer 

This paper presents a computer 
science program for the secondary 
school. The program covers both 
hardware and software areas. 
Course titles, descriptions, 
outlines, and suggested references 
are included. 


Today we have available inexpensive 
highly reliable computer systems; 
personal computer systems. The 
impact of -these -systems has- not 
been strongly felt yet, but it 
will be of ever increasing signif- 
icance as these systems become widely 
used. Right now we need to utilize 
these fantastic little machines in 
our educational institutions. They 
are not only capable of performing 
great amounts of mundane work at a 
fraction of the human cost in time 
and energy, they are a marvelous 
teaching resource, allowing for wide 
application in every subject area. 
They have enormous potential for 
providing students with the oppor- 
tunity for creative thinking, problem 
solving, and expanding awareness, 
also synthesizing knowledge. 

It has been several years since 
personal computer systems first 
appeared upon the scene, students 
in some classes have had the oppor- 
tunity to develop a basic system. 
But what now? Now that we have the 
computer system, what do we do with 
it? Unless we address ourselves to 


this question, schools' will become 
even more out of date, and the 
computer will become an idle piece 
of furniture in the same manner as 
the overhead projector, and the tape 

we must take advantage of the myriad 
possibilities computers offer. We 
need to develop and implement a 
scheme which will utilize computer 
systems in the schools. We need a 
computer curriculum that can be 
incorporated into the offerings of 
any school v-.'i!.-* =. mini ;um of dis- 
ruption, and which will provide 
students with the background and 
kmowledge needed in this area. The 
following courses, several of which 
have been implemented, are proposed 
to enable the schools to begin up- 
dating and incorporating the computer 
into the general curriculum. Com- 
puters are already in the mainstream 
of society, and should have that 
position and acceptance in the 
schools as well. These courses are 
'xtrrtrrely relevant and are vitrually 
needed by today's technologically 
extended student. The proposed 
curriculum would form a strong, 
viable program which would provide 
students with the knowledge and 
abilities needed at this time, to 
function in an increasingly 
computer oriented society. 

Each of the following courses could 
earn the student 1 semester of credit 
and have a maximum duration of one 
semester. Students might move 
through the sequence in a variety of 
ways, depending upon their interests 
and abilities. However, the first 
two courses are prerequisite to all 
others. Students may elect to 
challenge the contents of a course 
by examination, which is oral and 
written, and is administered by a 
computer science teacher. Successful 
challenges would move onto the next 
course in the sequence. It would be 
at the descretion of the principal 
whether to grant credit toward 
216 BOX 1 579, PALO ALTO CA 94302 

graduation for successful challenges. Computer Technology I and II 

Course descriptions are given in 
outline form only, space does not 
permit any detail. 

Computer Science I 

An introduction to the area of 
computer science which would 
include programming in BASIC, 
a study of computers in society, 
and an examination of the 
characteristics of computers. 

Objective: At the termination 
of the course each student will 
have developed, debugged, and 
run 10 programs that utilize the 
language BASIC, and the techniques 
of structured programming. Students 
will also have developed one project 
or program on a topic of their o%m 
choosing. All programs and projects 
will be presented to the instructor 
with a run and a listing of each 

Topic sequence: 
What are computers? 



human problems 
Communicating with your computer 




structured flowcharts 

programming in BASIC 

debugging techniques 
Developing your own programs 

determining if you can do it 

limiting the problem 

using your resources 

testing your hunch 

final production and publication 
Course project. 

Selected references: 
Albrecht, R., et. al. BASIC . 

New York: John Wiley and Sons. 

Hamming, C.L. Computers and 

Society . New York: McGraw-Hill 

Book Co., 1972. 
Kemeny, J. Man and the Computer . 

New York: Charles Scribner's 

Sons, 1972. 
McGowan, C, et. al. Top-Down 

Structured Programming Techniques . 

New York: Petrocelli/Charter. 

1973. ' 

This course presents the student 
with the opportunity of studying 
the computer and how it functions, 
maintenance, and troubleshooting 
techniques. The use of test 
equipment and component re- 
placement are also covered. 

Objective: At the end of this 
course each student will be 
able to determine if a system 
is functioning properly, if not 
to troubleshoot at the block 
level using appropriate test 
equipment and techniques to 
bring the system up once again. 
The student will also be able 
to identify and replace mal- 
functioning components at the 
board level. 

Topic sequence: 

What do computers do and how 

do they do it? 
Block diagram level of operation 
Electronic circuits 
Component level of operation 
Test equipment and it's use 
Troubleshooting strategy and 

Comprehensive examinations 



References : 

3yst^:n j.-ianu-.i ~ .'or the specific 

systems being used 
Test equipment manuals 
Handouts prepared by instructor 

Computer Technology III 

This course provides the student 
with the opportunity of exploring 
the construction of computers 
and related equipment. Students 
will be encouraged to construct 
from kits or components, items 
of use in a computer system. 

Objective: By the end of the 
course each student will have 
developed a computer device 
(anything related to comDuting) 
and demonstrated its proper 
functioning to the instructor 
and/or class. The student may 
use any resources within the 

Topic sequence: 



BOX 1579, PALO ALTO CA 94302 A 

Integrated circuits 
Theory and application of 

Techniques in electronics 
Design considerations in 

Use of design information 

(manufacturer's data) 


Articles from issues of: 
BYTE, Interface Age, 
Personal Computing, 
Kilobaud, and Popular 

Computer Science II 

A study of the application of 
decision tables, modularity, 
data structures, sorting, and 

Objective: Each student will 
be able to produce at least one 
computerized simulation by the 
end of the course. The topic 
or area of the simulation will 
be agreed upon by the instructor 
and student. A listing, and run 
of the simulation will be present- 
ed to the instructor as the course 

Topic sequence: 

Modularity in programming 

Decision tables 

Applications of decision tables 

Files and their uses 


Cassettes and discs 


Developing simulations 

what are they? 

characteristics of simulations 

designing simulations. 




Flores, I.(ed). Computer Sorting . 
Englewood Cliffs, N.J.: 
Prentice-Hall, 1969. 

McDaniel, H. (ed). Applications 
of Decision Tables . New York: 
Bradon/Systems Press, Inc., 

Maidment, R. and R.H. Bronstein. 
Simulation Games . Columbus , 
Ohio: Charles E. Merril Pub- 
lishing Co., 1973. 

Selected articles from: BYTE, 
Creative Computing, 


Interface Age, Dr. Dobb's 
Journal, and Kilobaud. 

Computer Applications in Eusiness 

A survey of the applications of 
computers to the field of 
business. Students will develop 
applications in any area of 
business: marketing, management, 
accounting, etc. 

The actual content of this course 
will vary from student to student. 
Specifications of the course 
content will be stated upon 
a student-teacher contract. 
(see Appendix B) 

Sample Objective: The student 
will develop and run a computer 
program that will process 
data in the area of accounting. 
The program will produce a 
full inventory, with . sales 
trends on a month by month basis, 
and include an automatic reorder 
feature. This project will 
be completed and submitted to 
the instructor before the 
termination of the course. 

References: It is suggested that 
the student be encouraged to 
survey the publications: BYTE, 
Interface Age, Creative Computing, 
etc. Also contact business persons 
in the community for information 
a~5" "to what" is actually needed. 

Computer Applications in Mathematics 

This course provides the student 
with the opportunity of exploring 
computer applications in the area 
of mathematics, including: number 
theory, geometry, numerical 
analysis, and statistics. 

Again, the actual content of this 
course will vary from student to 
student. A student-teacher 
contract will explicitly state 
what is to be accomplished. 

Sample Objective: The student 
will develop and produce a 
computerized study of the residues 
in mod 7 systems. The developed 
materials will be presented to the 
instructor at the end of the course. 

References: Mathematics texts, 
and programming manuals. Again 
the student should be encouraged 
to review the periodicals. 
218 BOX 1 579, PALO ALTO CA 94302 

Computer Applications in Natural 

This course covers applications of 
computer technology to the areas of 
physics, chemistry, biology, earth 
sciences, astronomy, biomedicine, 

The actual content of this course 
is determined by student-teacher 
contract. (Appendix B) 

Sample Objective: The student will 
produce a computer simulation in 
chemistry, specifically dealing 
with the determination of safe 
combinations of elements. The 
student will present to the 
instructor by the end of the 
course, all materials developed 
and tested, with a run and list- 
ing of each program. 

Sample Objective: The student 
will produce a simulation in the 
area of physics that demonstrates 
the behavior of projectiles in any 
atmosphere (Earth, Mars, Pluto, 
etc.). The student will present 
the simulation to the instructor 
with a run and listing before 
the end of the course. 

References: The references again 
will vary with the topic and the 
student. Standard texts in the 
subject areas, programming texts, 
and periodicals should be used 
as references. 

Computer Applications in Social Science 

This course provides the student 
with the opportunity of exploring 
applications in the social sciences 
including: psychology, history, 
economics, anthropology, sociology, 
and political science. 

The content of this course will 
be determined by the student and 
teacher through the use of a 
contract. (Appendix B) 

Sample Objective: In the area 
of political science the student 
will develop a computerized 
voting preference projection for 
a local election. The finished 
program should have been tried 
in an actual election situation. 
The materials, run and listing, 
with results of application, will 
be presented to the instructor 

at the end of the course. . 

Sample Objective: In the area of 
economics, the student will produce 
a computerized stock market analysis 
that will i; -ludo: h.'ghs, and lows 
for the year, and projections, 
based upon passed performance. 
The student will submit to the 
instructor a run and listing of 
the material developed during the 
course. This material must be 
submitted before the end of class. 

References: The materials needed 
to produce computer applications 
in this area or areas will vary. 
It is suggested that the student 
survey the regular texts, and 
periodicals in the area of his/her 

Independent Study in Computer 

This course will allow the student 
the opportunity of exploring 
applications in the areas of: 
music, art, danc^, literature, 
counsri-rg, etc. Areas not 
covered in any other course in 
the field of computer science. 

The content of this course is 
so variable that contracts 
between student and teacher 
are needed. (Appendix B) 

Sample Objective: The student 
will develop a computer program 
that will generate poetry, in 
the proper meter. The program 
will be presented to the instructor 
this includes a run and listing. 

Sample Objective: The student 
will produce a computer generated 
musical composition, and play 
the composition for the instructor. 
This will be accomplished before 
the end of the course. A run 
and listing of the program must 
also be presented to the teacher. 

References: In addition to the 
periodicals mentioned earlier, 
all available books, and persons 
familiar v/ith the area of study 
should be used as references. 

Note: Additional ideas and 
objectives can be found in 
Zeddies pages 61-67. 



BOX 1579, PALO ALTO CA 94302 

Computer Science xxl 

This course provides the student 
with the opportunity of studying 
assembly language, and machine 

Objective: The student will be 
able to produce at least five 
programs in any area, in assembly 
and/or machine code. The student 
will also develop an approved 
project and present it to the 
instructor with appropriate 
verification of correctness and 

Topic sequence: 

Assembly language 



Machine coding 



Programming in assembly 

Programming in machine 


Appropriate manuals from 
manufacturers of cpu chip. 
Instructor developed hand- 

Independent "Study in Computer 


The course allows the student to 
study computer languages such as 
FORTRAN, etc. 

The content of this course will 
have to be defined through the 
student- teacher contract* 
(Appendix B) 

Sample Objective: The student 
will develop, run and debug 
10 programs. The student will 
present to the instructor a 
listing and run for each of the 
10 programs. All programs are 
to be written in FORTRAN. 

References: The references 
needed in this course will 
,__y vary and therefore need 

to be stated on the contract. 
Include texts, and periodical 

Independent Explorations in 
Computer Science 

This course allows the student 
to explore the areas of trees, 
searching, compiler writing, 
and other advanced topics of 
interest to the student. 

Again the contents of this 
course will vary with the 
student. A contract should 
be used to explicitly state 
the material to be covered. 
(Appendix D) 

Sample Objective: The stu^-jr-- 1 
will develop an inter pre tor 
that will function within the 
limits of the available system. 
The completed project must be 
demonstrated for the instructor, 
and a written listing submitted 
at the time of demonstration. 

References: References will 
vary depending upon the topic 
of study, however the periodicals 
should be surveyed and it is 
suggested that you consider: 
Aho, A. V., et. al. The Design 
and Analysis of Computer 
Algorithms . Reading, MA: 
Addison-Wesley Publishing 
Co., 1976. 
Gries, D. Compiler Construction 
for Digital Computers . New 
York: John Wiley and Sons, 

1-971. - - - - - - 

Knuth, D.E. Fundamental Algorithms , 
Addison-Wesley Publishing, 

The proposed courses are not to be 
traditional in nature. They are 
quite individualized and are prob- 
ably best incorporated into the 
present curriculum in a multi- 
level configuration in which a 
classroom of students will be made 
up of several smaller groups of 
students, with each group pursuing 
a different course of study. This 
form of instruction allows a great 
deal of learning to occur between 
students, and enables the teacher to 
assume the role of a consultant, 
rather than the fountain of all 
knowledge. This also allows the 
students to observe the teacher 
in action as a learning being, 



BOX 1579. PALO ALTO CA 94302 

similar to the situations that 
would be found in a university or 
research center. In keeping with 
this format of organization, the 
teacher might want to consider the 
following: students are divided into 
groups of approximately 10, with 
each group having some beginning 
students, some advanced and some 
intermediate students working on 
projects of varying sophistication. 
The function of the groups is two 
fold: First to provide working 
groups with the opportunity to 
expand each student's understanding 
of the computer science area, and. 
second, to allow groups to meet in 
seminars where students can report 
to the group concerning their past 
two weeks work, their problems, and 
their accomplishments. In these 
seminars much information is present- 
ed and helpful suggestions are freely 
offered, the teacher also participates 
but does not dominate. 

The amount of equipment needed to 
implement this program would be 
minimal and could even be phased 
in over a period of two years. It 
is suggested that a minimum con- 
figuration for the entire program 
would include the following items. 

1. A multi-user system for 4-6 
terminals and at least BASIC, with 
discs, and cassette I/Os. 

2. A small system one user for 
assembly and machine coding courses. 
This system could also have a 
compiler/interpretor for wider 


3. A system with the capability of 
several languages. To be used in 
the more advanced courses and for 
heuristic explorations. 

It should be possible for all 
the systems to be configured for 
use as a computer network when 
this would be desired. 

4. A reference library that 
would contain books and materials 
covering : 

computers in society 
computer security 
computer hardtvare design 
manf acturer ' s data on CPUs 

and ICs 
structured programming 
computer applications in 

various disciplines 
compiler construction 
system development 
data handling. 

Also students would be in need of 
materials covering: 

numerical analysis 

simulation construction 

Boolean algebra. 

In addition to the above, subscriptions 
to the following periodicals should 
be maintained: Byte, Kilobaud, 
Creative Computing, People's Computer, 
Dr. Dobb's Journal, Personal Computing, 
and Popular Electronics. 

A diagram of the course sequence 
is' included in Appendix A. This 
diagram should clarify the sequence 
of courses as suggested in this 


The computer science curriculum 
proposed in this paper is but one 
possible configuration. This 
configuration presents the student 
with the opportunity of pursuing 
studies in the area of computer 
science at a depth beyond what many 
educators would believe possible. 
However, the work being done in the 
field of computer science at this 
time is basically the work of young 
people, and onemust not forget that 
young students are less hampered 
by the cultural inhibitions" which 
limit and stultify older generations. 
The experiences of the author have 
convinced him that the proposed 
curriculum is not only possible, it is 
a necessary change. 

The school system is guilty of sever ly 
limiting it's students, of failing to 
adequately prepare them with the kinds 
of knowledge and experiences needed 
in the face of the present day tech- 
nology. We must do a better job of 
meeting the challenges presented by an 
ever expanding technology. We must 
utilize all the resources available 
and provide our students with the kind 
of educational opportunities so vital 
to them in this day and age. 


Zeddies, M.L., et. al. Individualized 
Instruction for Gifted Students 
Using Computer Time-Share Systems . 
San Diego: San Diego Unified 
School District, 1974. 



BOX 1579, PALO ALTO CA 94302 




































to a; 
w c u 

o o c 

«d -P -h 
p <d u 

cp u to 

c o 

X -P 

&td P 

-p E 
c o 





•O -H 


LJ a 











W d> <L) -H 
U U U -P 
•H C C id 

w -p cd <d u 

W A3 'H -H -H 

0) £UUH 

c o to co a 

•h x: u 

w -p -p u < 

3 n) ia o 

I I I I <L> 

W W W W -P 
C C C C P 

o o o a 

H -H -H E 



+j 4-> -p -p d 

(d fd (d fd u 

u u u u 

•H -H -H -H C 

H H H H -rl 

a a a o, 

< < < < H 















Dates: begin_ 


bourse Objectives: 

Resources Required: 

Projects to be Produced: 

Evaluation. C.r i ter i a : . 

Contract Approved by: 

Student : 



Contract Completed: 
Faculty Signature:, 


Contracts to be made out 

in duplicate with each party 

receivma a codv, 



BOX 1579. PALO ALTO CA 94302 

(or, You Can Do It If You Try) 

Robert S. Jaquiss, Sr. f North Salem High School, Salem, Oregon 97301 


Now is the time for the (micro )computer to 
become widely accepted as a teaching tool in 
the subject area classrooms of mathematics, 
biology, chemistry, physics, business and 
social studies. 

This can be accomplished by the use of micro 
processors and the establishment on a high 
school level of a Computer Science Depart- 
ment which will implement a series of 
computer courses. These computer classes 
will form the basis to carry computer usage 
to the subject area classrooms. 

Teachers in the subject areas will need to 
be re-programmed to accept computer usage 
and to include computer usage in their 
lesson plans. 

There is need to make a 5-year plan for the 
acquisition of computer facilities, classes 
to be taught and specific programs to be 
implemented in specific classes. 

Statement of the Problem 

The computer has not been used to its full 
potential in education. 

The problem will be restated in various 
positive statements. 

The high cost is one reason computers have 
been used so little in education. Modern 
technology in the form of microcomputer 
systems is providing the capabilities to 
use the (micro )computer as a tool in the 
subject-area classroom at a more 
reasonable cost. The computer may be used 
as a teaching tool in many clever ways. 

Teacher T raining . Because this technology 
has come on the scene so rapidly, many 
teachers have received no instruction in the 
use of the computer as a teaching tool. In 
addition to not knowing how to use the 
computer as a teaching tool, these same 
teachers may fear the computer. 

The solution to these various statements of 
the problem seems simple. Just make use of 
available technology in education. 
This may be easier to say than to do. 

I have a plan 

Much software in BASIC already exists as 
the product of the last several years of 
research and computer use. 

The hardware to make use of that software 
is now available in the microprocessor 
based computer system. 

All we have to do is put the hardware and 
the software together in our school system. 

What A Micro Can Do 

In preparing this talk I wrote several pages 
about the capabilities of the microcomputer 
system. Perhaps those of us here already 
have a pretty good idea of what a micro- 
computer can do. On the other hand, perhaps 
some of you came here to the San Jose 
Computer Faire to find out what a micro can 
do. I guess that is why I came — to find 
out more about what is, and what will be 
available in the micro world. 

What did the Mini say to the Maxi Monster? 

"Anything you can do I can do better." 
And what does the Micro say to the Mini? 

"Anything you can do I can do better." 

If you haven't heard it before, then hear it 
from me. Practically anything you ever 
heard of a computer doing is being done by 
someone's microcomputer. In the exhibit 
area there are micros that talk, play music, 
draw pictures, make graphs and charts in 
color; some even compute. 
Did you see the walking talking computer? 

A micro can turn your lights on and off, 
lock your doors, call the fire department, 
keep your financial records and teach your 
children math or play games with them. 

A microcomputer system can be used 
effectively in the educational process in 
your school in math and science, English 
and social studies. All that is needed is 
to get the hardware and the software 

together and one more thing you. 

You will have to do the getting and putting 
because you are obviously the person in your 
school that is interest in computer use in 



BOX 1579. PALO ALTO CA 94302 

How to get hardware 

Establish a need . As you have noticed, I 
assume that some of you are teachers. OK 
Teach, try this. Walk into your princi- 
pal's office at 7:30 Monday morning, lean 
dramatically on his desk, look deep into 
his eyes and say: 

"The school district is failing in its 
obligation to teachers and students to 
provide modern equipment to use as tools in 
the teaching of academic classes and is 
therefore lessoning the value of the 
education provided to the students in our 

"The district may be considered negligent in 
not obligating teachers to seek additional 
training in the use of computer-based 
instructional methods. 

It is my guess that my now you have the 
principal's full attention. While the 
principal is trying to decide whether to 
fire you on the spot or give you a week's 
notice, you must quickly take a deep breath 
and continue- 

"Students are being denied the opportunity 
of experiencing hands-on use of modern 
equipment made available by the development 
of the microcomputer. 

The principal has probably gotten to his 
feet. You must say something soothing 
before he can get a word in edgewise. 
"Come let us reason together." And continue, 

"Our citizens of tomorrow are being deprived 
of the opportunity to use, see, and manipu- 
- late computer based - simulations-,- tutorials , 
and problem-solving techniques in the 
curriculum areas of biology, chemistry, 
physics, social studies, mathematics, 
business and even foreign language, because 
the equipment is not available for teacher 
and student use. 

At this point, the script calls for the 
principal to be seated again. With a smile, 
because he has decided you are not really 
dangerous, and because he has remembered 
plan X, he has an answer ready for you. 

The principal will say to you, "There just 
might be something to what you are saying 
Mr. Smith. Write up a proposal to take care 
of these problems. Justify all the parts of 
the proposal. Make out a list of equipment 
you will need, and the cost, and project it 
on a 5-year plan." 

Of course he will have to approve the 
proposal and pass the proposal up the line 
to program planning and evaluation. 

Justify the Need 

I suggest that you subscribe to a number of 
computer journals and read some books. 
One' article I like is "The Rhetoric of the 
Computer" by Barbara Marsh published in the 
Jan-Feb, 1978 issure of Creative Computing . 
On page 131 she says, 

"I believe people ought not emerge from 
schools at the mercy of what they see on 
television, what they read in the newspapers 
(if, infact, they read), or what a computer 
analysis tells them is the case. The 
television may define what are "the issues", 
and computer analyses may provide "the 
answers," but we must try to provide an 
education which leaves room for students to 
make their own evaluations and decisions. 
They should be able to assess the 
appropriateness of the computer mediation 
of the information dealt with, and have some 
idea of where to look to find what computers 
leave out. I think one way to make this 
more likely to happen is to have teachers 
who are able to think about computers as 
persuasive media whose output must be 

This article is one of a growing number of 
articles about computer literacy. 

Another article in the same issue, written 
by Cashman and Shelly with a very long 
title beginning, »"Hands-0n" And Fast Turn- 
around..." has some good thoughts. 
Cashman and Shelly say, "Teaching programm- 
ing without access to the machine is like 
teaching chemistry without access to a 
chemistry lab or teaching literature without 
reading a brook."" 

Cashman and Shelly are writing about using 
computers only to teach computer programming 
The reader should not be so narrow-minded. 
Computers should be used as a tool in the 
educational process . The points made in the 
article are very valid, but the scope should 
be enlarged to all users and not just 
computer programmers. 

ANY student should be able to use the 
computer in an interactive mode to the 
limits of his general ability . 

Cashman and Shelly provide a quote from a 
book by John Kemeny, Man And The Computer , 
published way back in 1972 by Charles 
Scribner's Sons, pages 80-81. Allow me to 
read this quotation to you. 

"I consider it imperative for the benefit 
of mankind that during the next decade 
computers become freely available to all 
colleges and universities in the United 



BOX 1579, PALO ALTO CA 94302 

I States and that most students before 
I graduating acquire a good understanding of 
| their use. Only if we manage to bring up a 
computer-educated generation will society 

have Modern CO*""te'' , s fully airailoKIs *■ ~ 

solve its serious problems. While computers 
alone cannot solve the problems of society, 
these problems are too complex to be solved 
without highly sophisticated use of comput- 
ers. I see three major bottlenecks that 
must be removed if this goal is to be 

"First, most university computation centers 
are still research-oriented. They are .... 
typically operated in a batch-processing 
mode with priorities given to a very small 
number of users who need a great deal of 
time. The philosophy of the university 
computation centers must be changed. 

"Second, college administrations do not yet 
appreciate the immense favorable impact 
that a good educational computation center 
can have on their institution. I would like 
to propose that by 1980 no college or 
university should be given full accredita- 
tion unless computer services are freely 
available to all students. Use of the 
computation center must be considered the 
exact analogue of the use of the library. 

"Finally, the implementation of this pro- 
gram for millions of students will take a 
great deal of money..." 

It is to quotations such as this that we 
may turn to for justification of the 
existance of a computer use program. 
Kemeny, and others, looked to the future. 
It is because of their efforts that the 
university computer centers have greatly 
improved since these words were written. 

Some colleges and universities now require 
a class in computer programming or 
statistics to graduate. At some institution 
consideration is being made to make a class 
in computer programming an entrance 

We might amend Kemeny' s words to include 
high schools: No high school should be 
given full accreditation unless computer 
services are freely available to all 


Kemeny had no way of knowing that the 
computer on a chip would become a reality 
so soon. While the implementation of a 
computer usage program in high school will 
take some money, the amount is many times 
less than what would have been required a 
few years ago, and the expectation is much 


The solution 

Simply get the hardware, the software, and 
£ou all together. You see, it all depends 

The school district should resolve to act 
with all possible intelligent purposeful 
planning to implement the use of the 
computer as a teaching tool. 

Computer Classes In High School 

Your school should have a Computer Literacy 
class to teach about computers and how to 
use them, how to use canned programs and 
how to do a little programming. Then the 
school should provide a Beginning Program- 
ming class for students who find they like 
computer programming. Finally, there should 
be an Advanced Computer Programming class. 
It is the advanced class that will run the 
systems, encourage others, write original 
programs, modify the library programs to 
work a little better and take the computer 
to subject area classrooms for 

Eventually one of the teachers will take the 
portable terminal to his classroom for a 
week to run simulations in biology class. 

This system is working for me and it will 
work for you. I also have students on an 
independent study program. 


Someone is going to ask about the money 
necessary to implement a new program. I am 
sure that the purchase of the first 16-mm 
projector for your school was a real event. 
Someone had to decide to purchase the first 
overhead projector. Believe me the money 
is there. If you can show the need for the 
program the money will come. The computer- 
use program cuts across the whole spectrum 
of studies in the subject area classrooms. 

The school district can afford to spend 
one-fourth of one per cent of its total 
operating budget for computer education 
and the program to use the computer as a 
tool in the subject area classroom. 

What a Microprocessor system can do . 

I was watching a program called NOODLE on 
a PET computer today. Fascinating. The 
little dot goes chasing around on the 
screen leaving a line and little rectangles 
behind it. With the proper persuasion I 
suppose the little dot could be taught to 
make graphs or maybe write one's name. 


BOX 1579, PALO ALTO CA 94302 

The computer, keyboard, CRT and a box to 
hold it all together costs only $800 with 
the additional memory option. A terminal 
usually cost more than that, 

I was going to try to define a micro- 
computer system and then I could talk 
about available software. But the system 
cannot do anything without 

Expectation Rises With Product Improvement 

There was a time when I was satisfied with 
a teleprinter throbbing away at 10 char- 
acters per second to slowly and noisily 
print out the results of my program. No more, 
I expect my printer terminal to print at 
least 30 characters per second and do it 

At one time I questioned the value of a CRT 

Zl^^z^-^^ ,.„*.*„ which leaves no written record but which is 

thinking first about the Huntington t x liked the 

simulations. The Huntington_I and Huntington ^IL?^ of ?hat printed 

Avai lable software . For software I was 

II simulations were written in primative 
BASIC by groups of knowledgable teachers for 
use in subject area studies. 

Each program is not very large. Most of 
these simulations run in very little memory 
The Huntington Programs are available from 
Digital Equipment Corporation. The 
Huntington I programs are available in 
listing form. The Huntington II programs 
are available on punched paper tape and 
come with three booklets, one each for the 
teacher, the student, and for additional 

The various computer journals have 
published a vast amount of computer soft- 
ware. Several computer companies have 
established user groups which maintain a 
library of programs available to members for 
the cost of reproducing the program. There 
is at least one private venture that is 
attempting to provide software to the user 
for a price, pay royalities to authors and 
make a profit. The hobby community 
exchanges programs so freely it may come- as 
a surprise that someone would sell programs. 
The big companies have been selling software 
for quite a while. If you purchase a mini- 
computer you will find it won't do much 
until you feed it several thousand dollars 
worth of software. 

Many books are available with program 
listings from simple games to General 

There exists a vast reserv oir of Computer 
software and computer knowledge relating to 
computer education in printed form as a 
result of several projects in computer 
education that have been undertaken on large 
time-sharing computers. 

Many of the computer programs thus devel- 
oped are available and can be executed on a 
microcomputer system. There is very little 
that a huge time-sharing computer can do 
that cannot be done on a microcomputer 
1 system. Those things that cannot be done on 
I a micro I didn't want to do anyway. 

security of that printed record to carry 
away with me. 

My students do not have this hangup. They 
make a printed record when handing xn 
assignments. More and more program output 
goes to the CRT in a form not amenable to 
character oriented printed output. So far 
I have insisted on a printer with each 
system so that output can be assigned to 
either the CRT or the printer. 

'What are the pieces that make up a minimum 
micro computer system? I suppose it would 
be good to have a check list of things we 
expect to have in a minimum system. When we 
purchase a system we will at least know if 
we do not have some of the items on the list. 

Everyone is not going to come up with the 
same list of expectations. It is easy to 
find microcomputer systems that do not have 
everything on the list. 

Minimum microcomputer sys tem for Computer 
Use in Education 

CPU board, name your favorite processor 

32K memory 

25-30 amp power supply 

extra slots for expansion 

CRT with ability to display memory 

locations, upper/lower case characters 
Matric printer terminal (keyboard, 30 cps) 

upper/lower case, adjustable paper size 
Extended BASIC 12-16K 
Operating system 

Assembler . 

Output may be directed to CRT or printer 

from within the BASIC language program 
Cassette storage system 

Such a system is available for a 
thousand dollars. 


Perhaps your list is different. No matter. 
The important thing is that you understand 
the limitations of the system you are 
thinking about purchasing. 



BOX 1579, PALO ALTO CA 94302 

Additional Options to expand the 
Minimum Microcomputer system for 
Computer Use in Education 

CRT with graphics capability 
24 or 48 lines of 80 characters 
Color CRT with graphics and plotting 
Matric printer with graphics and plotting 

bidirectional printing 

move paper forward and backward 

lower case descenders 
Letter quality printer/terminal 
Dual 8" floppy disk system with 

Extended disk BASIC with read/write 

files, chaining 
Paper tape I/O for use in transferring 

programs from one system to another and 

to read punched paper tape prepared off- 
Matrix printer/terminal with programmable 

character set for use in applications such 

as foreign language classes* 

More than one system, compatible with each 
other. Computer programs are written and 
developed on the development system and then 
transferred by cassette to the portable 
terminals that can be taken to any class- 

First Things First 

There are several thing you will have to do 
first before your school can make use of 
microcomputer systems in education. 

First you must have a computer to run. 

First you must be a programmer. 

First you must want to be a programmer. 

First you must be willing to put in lots and 

lots of extra hours at school and at home. 

First you must decide what kind of programs 

you want to run. 

First you must decide if you want an 

expandable or a closed system. 

First you must consider paper tape systems 

cassette systems, floppy disk systems, 

hard copy or CRT. 

First you must convince the administration 
of the need to use computers in education. 
That is where we were a while ago. The 
principal had just asked you to write a 
proposal to use microcomputers in the 
educational process. So first you write 
the proposal. 

One outline commonly used in proposals is: 
-Statement of the problem 
-The Proposal 

Curriculum changes 

Effects on students 

goals of the proposal 

teacher training 
- costs, and timeline for implementation 


Lets take some of the' problem statements 
and change them into goals to write into 
the proposal. 

The school district will provide 
one computer terminal for each- 

- 200 elementary students grades 1-6 

- 100 students in junior high grades 7-8 

- 50 students in senior high grades 9-12. 

The school district will- 

-provide the opportunity for teachers to 
take classes in computer uses in education 
for college credit. 

-encourage teachers to take these classes. 

-provide additional support in the form of 
books and magazines for the library. 

-direct the principals to procede forthwith 
to implement computer use in educationally 
sound ways. 

-recognize that every student has the right 
to be educated in Computer Literacy. 

-recognize that it is the obligation of the 
school district to provide the opportunity 
to the student to become computer literate 
at a level commensurate with his/her 
overall level of education. 

-thoughtfully consider that the school may 
be cheating those students who need 
computer literacy/programming to become 
effective citizens or college students. 

-admit that teaching using computers should 
occur whenever computers are an appropriate 
and educationally sound aid in the overall 
instructional process. 

Value to the Student 

Let us consider some of these same concepts 
in terms of those values and advantages 
that accrue to the student. 

The student is able to- 

-use applicable and available technology 
in acquiring an education. 

-investigate topics in his classes by using 
simulations that would not otherwise be 
possible because of 

time (heredity experiments in biology) 
danger to the student (heat, nuclear) 
mathematical drudgery (decimals). 

-become computer literate to suit the 
student's own prupose whether it be taking 
an elementary class about computers, or 
an advanced computer programming class. 

-use problem solving techniques by 
using prepared programs, or by 
writing original programs. 



BOX 1579, PALO ALTO CA 94302 

Value to the Student , cont. 

The student is able to- 

-learn how to program a computer in several 
languages* (Some of my students speak four 
dialects of BASIC, FOCAL, FORTRAN and two 
assembler languages.) 

-consider computer programming 
as a vocation 
as a recreational interest. 

-satisfy existing and/or proposed college 
entrance requirements. 

-have the experience of actually writing 
useful programs to be used in the education 
of other students. 

-do system programming. (My students have 
revised the operating system for several 
reasons by using machine language patches. 
They have written their own device handlers 
and disassemblers and are working on an 
operating system. 

The teachers of science, social studies, 
math and business will have the opportunity 
to use the computer to run prepared 
simulations in their classrooms. 

As a result 

-students will have the opportunity to run 
simulations, tutorials, and drill and 
practice programs relating to class work, 
either as an outside class assignment or 
just because they want to. 

-students will have the opportunity to play 
games on the computer. It is painful to 
some teachers to see students playing games 
on the computer but games represent problem 
solving and strategy in perhaps a different 
context. Many games reflect a learning 
situation. Sometimes there is an effort to 
disguise drill and practice as a game. 

Playing games very often leads the student 
into computer programming because the 
student wants to design his own game. 
Programming a computer to play a game is in 
itself, problem solving. 

Make a 5-year plan 

Realize at the beginning that next year you 
will have to make a new 5-year plan. 

Still, it is essential to plan where the 
program is going, or at least the desirable 
direction for it to go. 

Realistic goals change a great deal 
depending on available equipment, both 
hardware and software. 

I hope you will pardon some examples from 
my own situation to illustrate the poiM. 

In 197^ the computer at North was a PDP8I 
with *tK of memory. We had BASIC with 10 
or 12 statements and a user space of about 
1500 characters. We also had FOCAL which is 
a much better language. I have some great 
student programs written in FOCAL. There 
are a number of games. The most ambitious 
program is one that will solve 9 
simultaneous linear equations in 9 unknowns. 
We did not do much in the way of program 

The first computer I programmed was an 
early IBM monster that required a suite of 
airconditioned rooms. In 1957 we 
p rogrammed in machine language and 
optimized the program by strategic placement 
of variables in storage locations on the *fK 
drum memory. I was proud of my program 
that successfully added two numbers. 

With this kind of equipment goals may be 

I still teach my Computer Literacy classes 
to use machine language to add and multiply 
two numbers, but only to demonstrate how 
the computer works. 

In September, 197^* my 5-„ r ear plan was to 
teach computer programming on the PDP8l in 
machine language and FOCAL, and to teach 
time-sharing BASIC using a teletype connect- 
ed to Oregon State University. The 
addition of a highspeed paper tape reader 
made it reasonable to use the assembler on 
the PDP8I. 

In February, 1975. I ordered an Altair 
computer from 'KITS" "for" my personal use. 
The world was about to come unglued. 

In September, 197^» I had no reason to hope 
or even expect to hope that I would be able 
to do anything useful in computer 
programming classes. 

The time-sharing Teletype was connected by 
leased long-distance telephone line and was 
costing a large amount of money. There was 
no educational program library on the time- 
sharing computer, although there was a large 
technical mathematics library. This was the 
state of the art at that time. 

In my summer school classes I heard about 
marvelous concepts like PLATO and 
HUNTINGTON simulations. I think I remember 
saying, with some bitterness, "What good 
is all this going to do me?" 

I ran many decks of cards through the IBM 
360 on Free WATFIV. FORTRAN on cards and 
the slow turn-around-time really did not 
excite me very much. 



BOX 1579, PALO ALTO CA 94302 

Expectation rises with better equipment 

So in September, 197^, I really had low 
expectations. I did have a $5000 per year 
budget for the time-sharing terminal. 
When Digital Equipment Corporation 
announced the CLASSIC in December, 197*t, 
I was ready. The time-sharing was 
cancelled and the money thus saved was 
used to purchase the CLASSIC on a 5-year 
lease-purchase plan. 

The Digital Equipment Corporation's 
CLASSIC computer has a PDP8A CPU, l6K of 
12-bit core memory (effectively 32K char- 
acters), dual floppy disk (256,000 char- 
acters per disk) and a keyboard CRT (12 
lines of 72 characters). We have since 
added a print-only DECwriter for hard copy 
output, and a Teletype ASR33 for punched 
paper tape I/O to the system. We also 
purchased software including OS/8, CLASSIC 
BASIC, FORTRAN IV, and PAL8 assembler. 
One of the students wrote an 8K FOCAL 
language processor for the CLASSIC. 

.e have all of the Huntington II simulations 
and "101 BASIC Games" on disk. In inkprint 
we have all of the Huntington I programs 
and several publications of application 

All this is costing us less than $250 
per month. 

Now the five-year plan has to be 
completely re-written. Expectation has 
gone up because of available hardware. 
In the meantime we have added 2 ASR33 
Teletypes for off-line punched paper tape 
preparation of programs. 

Our situation has changed completely. 
We now have a large software resource 
but only 2 terminals. Students are using 
the CLASSIC and the 8l all day long. 
After a brief time of playing games and 
running the Huntington simulations these 
students are interested only in creating 
their own programming masterpieces. 
The goal of using the computer in subject 
area classes is not being implemented 
because the computers are already being 
used fulltime for a worthy purpose. 

For part of the spring term we were able 
to rent a timesharing terminal connected 
to the HP2000F at Willamette University 
in Salem. This was used briefly in junior 
high classes and in biology classes. Money 
came from a grant from Oregon Mathematics 
Education Council (OMSC) which was in turn 
funded by the National Science Foundation. 
This experience proved to me the validity 
of the concept of computer use in subject 
area classrooms. 

In November 1976 we ordered a SOL-20 with 
dual floppy disks. This purchase was made 
possible by a matching grant from OMEC. 
When the unit had not been delivered by 
April, 1977, the order was cancelled and a 
microcomputer system was designed from on- 
the-shelf items at the local Byte Shop. 
This system was delivered in two weeks and 
was used the last 5 weeks of school. 
Several subject-area demonstrations were 
made during the last weeks of school. 

I am telling you about the history of our 
local situation to provide an example of 
how expectations change with the kind of 
equipment available to be used. 

In September, 1977, our computer #3 
consisted of an IMSAI mainframe, Cromenco 
Z-80 CPU board, 2kK. memory, VDM-1, Cutts 
Cassette interface. Byte Saver board with 
modified TDL system Monitor, TDL BASIC, 
CRT Monitor, DECwriter LA36, ASR33 Teletype 
and a cassette player. This is a powerful 
and versatile system. Either keyboard can 
be used as the console (not both). Output 
may be directed to either the CRT monitor 
or the printer associated with the keyboard 
being used, under software control in 
several ways: as a monitor assignment, 
by the 'switch* command in BASIC, or by 
BASIC statements of PRINT and LPRINT. 
Programs may be input or saved on paper tape 
or cassette. This system cost $3800, plus 
the Teletype purchased earlier, plus the 
DECwriter. Many computer stores now have 
financing available. You should be able to 
purchase such a system for $186 per month 
for 5 years. 

Let's review: 

In September, 197^, available equipment 
included Computer #1, a PDPSl (4K) with an 
ASR33 Teletype, and a timesharing terminal 
connected by telephone to a university *t0 
miles away. 

By September, 1975, the timesharing had been 
dropped and a higspeed paper tape reader 
added to the PDPSl. A CLASSIC had been 
ordered but was not delivered until 

By September, 1976, additions included 
Computer #2, the CLASSIC with DECwriter, 
and large amounts of program material on 
disk or printed. 

By September, 1977, we had added computer 
#3, the IMSAI-TDL system described above. 

I hope you don't think that is the end of 
the story. It isn't. We need more 
terminals and more options for a growing 



BOX 1579, PALO ALTO CA 94302 

If You Want Equipment, ask for it 

I am happy to report that my school has 
just approved the immediate purchase of 2 
new computer systems with 19-inch color 
CRTs, color graphics and plotting capability 
One of these systems will be available at 
all times for subject-area classroom use. 
The other system will be used for program 
development most of the time. 

I think I have been able to get this equip- 
ment because I have kept up a continual 
barrage of computer proposals, letters, 
and copies of articles from computer 
magazines to teachers, the administration 
and to the school board. 

I have had a lot of help from OMEC, from 
people I met at the University of Oregon, 
from Willamette University in Salem, and 
from students and faculty at North. I 
subscribe to many of the computer magazines 
and I put in a lot of hours. I would like 
to think we have the beginnings of a good 
computer education program at North Salem 
High School. 

I am not trying to get my own back patted. 

My message is just this: 

-the software is available 

-the hardware is available, 

-you can do it, too. 

Why are you just sitting there? 

If you want equipment, ask for it. 
Ask again, and again. 

Design a program for your school that looks 
to the .future. What is. desirable now? 
—next year? —in five years? 

Plan classes in Computer Literacy, 
Beginning Computer Programming, 
Advanced Computer Programming. 
U se BASIC, Assembler, FORTRAN, PASCAL. 

Plan for computer usage in the subject area 
classroom. Read those magazines, write 
another proposal. Attend that conference, 
write another proposal. Visit another 
school, write another proposal, and keep 
hammering away. 

North Salem High School cannot compete with 
Lawrence Hall of Science the first year, 
but if we go back into history far enough 
we will find there was a day when Lawrence 
Hall got its first computer terminal. 
Lawrence Hall just got a headstart. 

I believe that the computer should be used 
as a teaching tool in sub.ject a rea class- 
rooms . A class of advanced programming 
students can be a great help in implementing 
this program. We are starting with the 
Huntington simulation programs and re- 
writing them with improvements that are 
possible with a better BASIC. We will 
answer 'y es ' or ' no ' instead of '1' or '2'. 
Now we will modify the programs to take 
advantage of the possibilities of a CRT 
that can display kS lines of characters in 
any color and that has graphics and plotting 

We may get programs from books and/or 
magazines but most of the time we just 
write our own. 

Write another proposal, Ask the librarian 
to order computer books and magazines. 
Answer the ads in the computer magazines by 
using the bingo cards. 

You do not have to have a PDP-10 or an 
HP 2000 f or an IBM 360 to run worthwhile 
programs. Many BASIC programs will run on 
a stand-alone microcomputer that can be 
taken to any classroom. 

You can do it if you try. 



BOX 1579, PALO ALTO CA 94302 

San Jose Computer Faire, Jaquiss, Computer U6e in education HANDOUT 

1* Is easy for the user to use; teacher, novice or programmer. 

2. Provides for a number of programming languages, 

3. Permits the user to access a large variety of on-line canned programs 
for simulation, problem solving or computer literacy. 

k. Gives the user a choice of using CRT or teleprinter terminal. Output 
can be directed to either the CRT or printer from within the (BASIC) 
program while it is being RUN. 

5. Is flexible so that terminals may be plugged in in any classroom for 
use in the subject area classroom. Allows the use of a large 25-inch 
CRT monitor, printing terminal, plotter or graphics terminal, or is 

a portable system that can be moved to the desired classroom. 

6. Offers the user the options of graphics, a plotter, color CRT, line 
printer, CRT, or printing terminal either matrix or letter quality. 

7. Allows the user to have his own user number and secret password, and 
to have protected files in hi» own private library, or else his own 
floppy disks. 

8. Has sufficient memory so that each user has adequate user space. 

9. Provides for easy transportation of programs to and from the system. 
Locally produced programs can be provided to other computer centers. 
Software secured from outside sources can be put onto the system via 
magtape, punched paper tape, cassette tape, floppy disk. In the case 
of several micro systems programs should be easily moved from one 
microcomputer to another. 

10. Is expandable. Either upward compatible as in the PDP-11 series or 
the microcomputer system has extra slots available installed. 

11. Provides the user a method to save his programs outside the computer 
(punched paper tape, tape cassette, floppy disk). 

12. Provides system backup in case the system crashes. Systems do crash. 
In some systems the entire disk is copied onto mag tape at regular 
intervals, tfith a floppy disk system the user can make his own 
back-up disks. 

I forsee intelligent terminal networks coming to the aid of microcomputer 
systems with each intelligent terminal being able to access a large 
intelligent disk data base. The intelligent terminal will go to the disk 
to get whichever language processor is desired and copy it into memory. 
Then the intelligent terminal will access the master disk again to load 
(or save) programs. The intelligent terminal will be able to execute the 
desired program in the selected language. A user will be able to select 
a CRT terminal or a printing terminal from which he can control a high 
speed line printer, reader punch, cassette tape or floppy disk. Schools 
will have a jack in every room so that any terminal may be plugged in. 
The system will provide color CRTs with graphics and plotting and a color 
camera copier. There will be a system plotter and a line printer capable 
of 200 dots per inch resolution. 

I would like to teach my microcomputer to groan when the student makes a 
mistake and maybe play "On North High" when the program runs. 


COPYRIGHT 1978 Don Black 


DON BLACK, Formerly Director of Computer Activities, the Learning Farm 

now with 

3304 Pico Boulevard 

Santa Monica, California 90405 

(213) 450-2060 

The Teachers Requirements for an Educational 

This presentation may be functionally divided 
into two sections: Hardware and Software. 


Packaging . The most important part of 
the system is the student, so I will start 
with what the student sees first, the Package. 

It is well to have the system in one or 
two self contained modules. A jumble of wires 
and boxes can be somewhat intimidating. A 
clean, simple unimposing package, without too 
many lights and buttons on the other hand, 
invites exploration. Sophisticated Package . 

CRT ys Har dc opy . CRT's are nice, quiet, 
fast, but they cannot provide Hardcopy. 
After I have spent an hour being creative, I 
would like to have something to show for it. 
If the presentation device provides Hardcopy, 

-The student can take it home. 

-The student and teacher have a record of 
the student's progress (we compile a portfolio 
of student progress in the form of this 
annotated output) . 

-The author has some feedback on course- 
ware effectiveness. 

If I have a choice between a CRT and a TTY, 
I choose the TTY for the reasons outlined 
above . Hardcopy . 

Graphics . The only reason I would choose 
a CRT over a Hardcopy device is for a graphic 
capability coupled with a medium speed 
data rate. The graphics need not be elaborate, 
just easy to use. The graphic capability need 
be no more sophisticated than drawing a point 
(or i/4" square) at x,y. For the author to 

to be able to specify a vertical or horizontal 
line (as with The Apple II system) would be 
better, but we are venturing into software, or 
at least firmware. 

Yet there is still very little that can be 
done with a TTY. This final decision may very 
well be a matter of taste. I still prefer the 
lowest common denominator, the TTY. Graphics 
is Fun, But Unnecessary . 

Memory . All the courseware I have written 
has resided happily in 6K bytes. To this must 
be added the software and firmware. Lets give 
software 4K (the author language itself) and 
firmware 2K (monitor, I/O, integer multiply /di- 
vide, and utilities) so our system requires 
10K RAM and 2K ROM. If we went to a nice round 
number like 12K, we would have room for some fun 
added capabilities. 10 - 12K RAM and 2K ROM . 

Storage . Our system will require a form of 
storage that will allow the student to painlessly 
load a course without any training. I imagine a 
cassette tape with a big green 'LOAD' button. 
Punched Paper Tape is too slow and troublesome. 
It requires both patience and understanding, 
somewhat limited commodities. A floppy disk 
might meet our requirements, but a cassette tape 
could do so also, and it is still cheaper and 
more familiar. We all look forward to the day 
of the bubble memory cartridge, which could also 
allow virtual storage capabilities, yet the 
cassette is cost-effective today. Cassette 
(Until Bubble Memory is Cost-Ef f ective ) . 

A cassette would allow easy transportability 
of courseware, and with cassette standardization, 

So, out of this discussion comes: 

1) Sophisticated Package 

2) Hardcopy 



BOX 1579, PALO ALTO CA 94302 

3) Graphics is a Luxury 

4) 10 - 12K RAM with 2K ROM 

5) Cassette Storage 

More on Cassettes . We can get by on the 
audio cassette toys, but in increasing orders 
of sophistication, allow the author to: 

a) turn the cassette on and off program- 

b) select read or write mode (say play or 

c) search a tape for a specific record. 

d) select a high speed search mode (read) . 

e) play back a pre-recorded message (audio) 

At level 'd' we have the capability of 
virtual storage, if the response time on the 
cassette transport is such that it will allow 
the author 1/4 - 1/2 second response time for 
student interaction. Once we reach the 'e' 
stage, we shall truly have an audio-video de- 

I initially prefaced my remarks by sayinp; 
that the student is the most important element 
of the system. Yet, in this field we are all 
students. The obvious corollary is the the 
User is the most important element of the 

The first requirement of our software is 
a clear, simple and direct syntax. There 
should be no need to re-educate a specialist. 
The function of the language elements should 
be clear. The variable names should be self- 
documenting (i.e. multi-character). The 
syntax should be easy to parse by a micro- 
computer based system. 

Let me introduce the -PILOT language at 
this point. This 'author' language, as 
developed by Dr. John Starkweather, meets our 
initial requirements admirably. I have taught 
9 year old children, with no computer experience 
how to program in less than an hour. It is no 
exaggeration to say that one may become profi- 
cient in the language within a few hours. 

I have selected PILOT for its unique 
adaptability to microprocessor implementations. 
There is a brief description of the language at 
the end of this paper. 

Box 354, Palo Alto, California. Let me plug 
my interpretor system written in FORTRAN and 
funning on Honeywell 6600, Univac 70/7 and 
Burroughs 6700 systems. This is a program 
product from the EDUTECH Project, Box 1023, 
Encinitas, California 92024. 

The PILOT syntax is as follows: 

[label] [operation code] [condition] : [object] 

The operation code for the core language is 
one character: 

'T' - Type, 'A' - Answer, 'M' - Match, 

'J' - Jump, 'U' - Use, 'E' - End, 'C' - Compute, 

'R' - Remark. This will come up later. 

Document at ion . The next requirement for any 
software is clear, simple documentation. This 
requirement is also first chronologically. At 
least outline the documentation before you begin 
designing the software. The User requires, no 
matter what her/his level of sophistication, 
a clear concise description of each element of 
the software tool. Be sure your assumptions as 
to the requirements of the user are clear in 
your own mind and be consistent throughout your 
presentation. Except for audiences like this, 
I usually assume a sixth grade reading level, 
that is a high school graduate. 

Include a table of contents, an index, and 
references to other sections of text that bear 
on the subject at hand. 

For example, when I say a description of eact 
element, I mean specify that in an assignment 
statement (or LET or computational) that evalua- 
tion proceeds from left to right and that 
exponentiation is evaluated before multiplicatior 
division, addition or subtraction. But don't 
say it like that. More like this: 

"Arithmetic proceeds from left to right, 


(and include examples). "X = 5+6-3" "6 is 
added to 5 to give 11, then 3 is subtracted from 
11 to give 8. 8 is then saved is 'X'". But, 
powers of number are computed before multiplica- 
tion or division. Multiplication and division 
are computed before addition or subtraction. 
For example, in "X = 10-2*3" "2 times 3 is 
computed first to give 6. 6 is then subtracted 
from 10 to give 4. 'X' now has the value 4". 

PILOT is available for the INTEL 8080 chip Use illustrations, too. "A picture is 
through the National Library of Medicine for worth..." and so on. 
free (public domain) . There are implementations 
available for most mini and maxi computer systems 
written in BASIC, FORTRAN and even PL/1 through 



BOX 1579, PALO ALTO CA 94302 

Order Of 




** Exponentiation 


*,/ Multiplication 


+,- Addition 

(Note that I am assuming the reader understands 
+,-,*,/>** operations.) 

Arithmetic (Assignment or Compute State- 
ment) . Integer arithmetic will serve our needs 
just fine. In fact, floating point arithmetic 
will create some problems due to rounding error 
that the student and teacher will not be pre- 
pared for at this stage (infinite loops due to 
improper testing) . 

There are, of course, many applications 
where floating point is necessary. An optiumum 
implementation would be a combination of float- 
ing point and integer (or fixed point) arithme- 
tic as in Fortran or PL/1. 

The arithmetic operations necessary are: 
addition, subtraction, multiplication and 
division. Exponentiation is good, but not 
really necessary. There are algorithms 
available for performing higher level functions 
given a set of lower level primatives: multi- 
plication from addition, exponentiation from 
multiplication, the 'n'th root given the 
arithmetic primatives. (Note - Include some 
capability for testing for a non-numeric input 
rather than bombing out a program) . 

String Manipulation . This capability is 
a necessity for any teaching algorithm more 
sophisticated than arithmetic or multiple 
guess (choice) . 

The primatives necessary are: 

1) Input a string (assign to a variable) 
lb) Assign a string (" ") 

2) Index (where in string 'B' does string 
'A' occur, if it does?) 

3) Substring (Extract from string 'A' 
a substring beginning with location 'B' and 
ending with location 'C'.) 

4) Concatenate (Form a third string by 

I placing string 'B 5 at the end of string 'A'). 
5) Output a string 


6) Conversion (Convert string 'A' to its 
numerical equivalent, or convert a number to 
its character representation) . 

7) A 'replace' or substitute operation 
would be nice, but it may be woven from the 
above primatives. 

Relational Operations 

1) Equal or Not-Equal and Greater Than 

2) Greater Than, Less Than, Greater Than 
or Equal To, Less Than or Equal To, Equal To, 
Not Equal To. 

3) For strings, a byte by byte compare 
using the EBCDIC or ASCII colating sequence up 
to the length of the shorter string. If strings 
are equal, then the longest string has a greater 
value . 

4) A value of 1 for TRUE, or a value of 
for FALSE. 

Logical Operations . Although these prima- 
tives are not necessary, they are handy. If 
you include the above two capabilities, these 
might as well be included also. This capability 
also lends itself well to structured programming 

1) Logical Or 

2) Logical And 

3) Logical Not (Unary) 

4) Logical Exclusive Or (While not a 
primative, it is useful). 

It might be interesting to use more contem- 
porary logical primaries such as: 

1) NAND (Not-And) 

2) NOR (Not-Or) 

3) Invert (Logical Not) 

For evaluation purposes, a value greater 
than zero is TRUE, while a value less than or 
equal to zero is FALSE. A Result evaluated to 
TRUE is set to one, while a result evaluated to 
FALSE is set to zero (state this in your docu- 
mentation) . 

Language Elements . The basic language 
functions that a teacher's language must perform 

- Present Information 
Get Response 

- Evaluate (Compare response to expected 
result. ) 

- Decision (Branch or conditional executioi 
based on result of analysis.) 

234 BOX 1 579, PALO ALTO CA 94302 

Present Information . We need an output 
instruction that allows the author to display 
text, values of variables, and primative 
graphics if available (clear screen, move cursor), 
An added capability would be to evaluate and 
display the result of variable manipulations. 

The PILOT language 'T: n statement meets most 
of these requirements. 


GET Response . The teacher needs to input 
the student response for evaluation purposes. 
The language requires the capability to identify 
the student response and save it for later evalu- 
ation. It is therefore necessary to assign 
input to variables. We require both integer 
variable and string variable input assignments. 

The PILOT 'A:* input statement is our 
example : 

A:#RESULT Would assign a numeric response 
(integer) to the variable 

A:$NAME Would assign the input string 
to $NAME. 

It is also necessary that a non-numeric 
response, when a numeric response is expected, 
does not 'bomb-out' the program. Allow the 
programmer the ability to test for a valid 
number. There is a conditional in PILOT that 
is set or reset if the last inputted string 
can or cannot be evaluated as an integer, 



(B is set if the number is Bad, G is set if 
the number is Good. The second and third lines 
of PILOT code are executed only if the 'B* con- 
dition is set, i.e. last inputted string is not 

Evaluate . The simplest form of evaluation, 
and the most popular, is to compare a student 
response with an expected response. A straight 
forward character by character compare (match) 
is the obvious solution: Does the student 
input contain the character string? The 
PILOT implementation is: 

M:string 1, string 2,..., string n 

If the last inputted string contains the 
characters 'string 1* or the characters 
'string 2* anywhere in its text, then the match 
is successful. Success is indicated by setting 



a condition flag "YES" for success or "NO" 
for not a successful match. This condition 
can be used in the same manner as 'B' men- 
tioned earlier. For example: 





If the characters string PILOT INFORMA- 
TION EXCHANGE or P. I.E. OR PIE occured in 
the student response then the student would 
get the reinforcing message "RIGHT."'. For 
example, if the student responded I LIKE PIE. 
The system would respond RIGHT! 

We also require the capability to manip- 
ulate inputted strings and numbers. For 
example, should an English teacher wish to 
use a transformational grammar approach in an 
English course, it will be necessary to 
evaluate an input string for syntactic vali- 




Rather than test for all syntactically 
valid responses, we would like to substitute 
the proper part of speech for each word and 
evaluate the grammatical elements. Let us 
assume a REPLACE instruction: 

REP: pattern, string, object 

where all occurances of pattern in object are 
replaced by string: 






BOX 1579, PALO ALTO CA 94302 

As evaluation proceeds, the teacher has 
the option of displaying the parts-of-speech 
and evaluated grammatical elements for the 


the black house jumped over the brown 

art adj noun verb prep art adj 

n-phrase verb prep phrase 

dog front the cat 
noun prep art noun 

subject predicate 



Similar evaluation may occur with arith- 
metic and math word problems. 

Another implementation of this capability 
is in an arithmetic assignment statement with 
string operations as with BASIC or PL/1. That 
is the method of implementation used in our 
PILOT Interpretor/Editor system (PI/ES) , 

Using the index, extract substring, and 
contatenation operations, I have written a 
subroutine to replace a given word with its 
part-of-speech, or a grammatical element with 
its term. 

Referring back to our requirement for a 
simple direct syntax, I don't think our PIES 

system meets this requirement in the area of 
string manipulation, although it is the best 
implementation I have seen. Something on 
the order of the REP instruction seems more 
straight forward. As they say in the education 
industry, I will leave this problem as an exer- 
cise for the student. Let me know if you 
solve it. 

Decision . The fourth elementary require- 
ment is the ability to make a decision based 
on our previous evaluation. The buzz word 
here is "conditional execution". Most language 
implementations have a conditional branch 

The syntax of this capability is far 
reaching, perhaps even devious. The syntax of 
this implementation defines the structure of 
the programs that are written with the language. 

Those of us who are teaching the art of 
computer programming look upon the BASIC syntax 
with horror. (Ever try to debug somebody elses 
4000 statement BASIC program?) A syntax such 
as PL/1 or PL/M however, is a joy. The 
IF A THEN DO ; . . . END ; sequence lends itself 
exquisitely to structured programming. 

Although PILOT is not a structured language, 
it has a syntax that is unusual and easy to 
explain: Every statement may be conditionally 

The condition may be in the form of a flag 
that is set by a previous statement: G/B set 
by the last A:; or Y/N set by the last M: , 
or the condition may be in the form of an 
arithmetic evaluation: 

J (A*B=5) :*L00P 

The language may be structured by adding 
another instruction that will group statements 
into a block. This may be considered grammati- 
cally as a pair of parenthesis: 



IF (I) 100, 200, 300 

J J. '. "ljABEu 

for BASIC 
for PILOT 


The spelling is unimportant, it could be 
called PROBLEM, or BEGIN, or SECTION, or GROUP, 
etc. The critical function is the conditional 
execution of the entire block. If the condition 
associated with the BLOCK statement is false, 
then execution of the entire block should be 
skipped (down to corresponding ENDB statement) . 



Let's talk about subroutines. A subroutine 
reference is a branch or jump to another location 
in the program code and a return to the next 
location following the subroutine reference. 
PILOT calls this a U: for Use. The subroutine 
performs a programmer defined function that is 
used frequently by more than one section of the 



BOX 1579, PALO ALTO CA 94302 

program. While both PILOT and BASIC include 
this capability, there is one shortcoming 
shared by both languages: There is no direct 
way to send variables to the routines. 
Notice the differences: 


CALL SUB (argl,arg2, "STRING", arg4) 

LET Al=argl 

LET A2=arg2 


LET A4=arg4 

GOSUB 1000 


I would like to suggest an extension to the 
U: instruction so that it may allow an 
Argument List: 

U : *SUB , argl , arg2 , "STRING" , arg4 

MISCELLANEOUS. Other capabilities that 
have been suggested I will mention in passing: 


Jump to the Nth listed address depending on 
the value of //VALUE. Similar to the FORTRAN 
computed GO TO: 

GO TO (100, 200, 300, 400), JVAL 
or in BASIC: 

IF V=l THEN GO TO 100 
IF V=2 THEN GO TO 200 

I would like to propose a select statement 
SEL that would select the Nth element from a 

SEL:#n, target, elementl,element2,element3. . . 
Usage would be as follows: 







This statement will allow the BRANCH 
capability and a pseudo dimensioned variabel 

External Storage. A File input or output 
capability is limited by the hardware. 
Assuming the lowest common denominator, Audio 
Cassette or PPT, we are allowed to READ. 

So the first capability we wish to im- 
plement is a sequential file READ, i.e. READ 
next record and assign to variable (s). 

This hardware capability also will allow 
us to CHAIN to another teaching program: 




CHAIN Y:*LESSON2,argl,arg2, . . . ,argn 

This instruction is a combination Load 
and Run. *LESS0N2 would be located, loaded 
from the tape into memory, and executed using 
the specified argument list. 

With this, instruction, we have the 
capability of n 6K byte modules or chapters 
of lessons. 

A File WRITE instruction would allow the 
author to store the results of student 
performance for later analysis (grading, 
student progress, course effectiveness, 'etc.) . 

READ : %f ilename , $variablel , #variable2 
WRITE : %f ilename , $variablel , #variable2 

A syntax including a filename (^filename) 
will allow upward compatability with systems 
that have more sophisticated I/O devices 
diskette, digital cassette, etc.). We might 
even include a syntax for Indexed files: 

READ : %f ilename * key , $variablel , #var iable2 

where "key" (following the apostrophe (')) is 
an integer from to 255 that refers to the 
Nth record of the specified file %f ilename. 



BOX 1579, PALO ALTO CA 94302 


In summary, our language consists of: 

1) Documentation 

2) Arithmetic Computation 

3) String Manipulation 

4) Logical Operations 

5) Relational Operators 

6) Output to TTY 

7) Input from TTY and assign to variable 

8) Compare (match) student response to 
expected response 

9) Evaluate response 

10) Conditional Execution 

11) Structured Syntax 

12) File READ /WRITE 

13) CHAIN to program using argument list 

If there is an OEM who can meet these hardware/ 
software requirements for less than $1000 please 
let me know, I would like a few. 


Pilot La ngua ge Syntax 

[*label] [op-code] [condition] : [object] 

[*label] a 1-19 character unique string, (optional) 

[op-code] a PILOT instruction as follows: 

T Type out to the terminal the string 
contained in [object]. Interpret a 
string prefaced by # as a numeric 
variable. Interpret a string prefaced 
by $ as a string variable, otherwise 
type out the object as is. 

A Accept an Answer from the terminal and 
optionally assign the input to the 
variable contained in [object]. SET/ 
RESET G/B conditions. 

M Match (compare) the last inputted response 
with the string optionally delimited by 
comas (,) in [object]. SET/RESET Y/N 

J Jump to the label contained in [object]. 

U Use [object] as a subroutine (save this 
address for return) . 

E End subroutine (no [object]). Return 
to instruction following last U state- 
ment . 

C Compute, [object] is of the form 
[variable] - [arithmetic expression] 
Evaluate [arithmetic expression] and 
assign result to [variable]. 

R Remark. Internal documentation, 
ignore [object]. 

Multi character [ op-code ]'s are used 
for system dependent or user defined in- 

REP [condition] : [pattern] , [new pat- 
tern] and update [old string] when 

CHAIN [condition] : %[ program-name ] , 
[argument 1], ..., [argument n] 
Load and execute (with no return) 
[program-name] using argument list 
[argument 1] through [argument n] 
SEL [condition] : [N] , [target] , 
[element 1] , [element 2] , . . . , 
[element n] 

Select the Nth element from the list 
and assign to [target], [condition] 
is of the form: [Y/N] [G/B] 
[ (arithmetic expression) ] 

Y is true if the last M match state- 
ment was successful. N is complement. 

G is true if the last inputted string 
could be evaluated as an integer. 

B is G's complement. 

(arithmetic expression) is evaluated 
as True ± f the result is greater than zero -;■ 
otherwise it is False. 

[type code] [variable name] 

[type code] is: '*' for labels, '#' 
for integers, '$' for string variables. 

[variable name] is 1-19 characters, 
first character a letter. 



COPYRIGHT 1978 Don Black 

BOX 1579, PALO ALTO CA94302 


Michael R. Levy 


70 Boston Post Road 

Way! and, MA 01778 


The "Hobby" manufacturers are turning more 
and more to the low end small business market in 
order to take advantage of this fast growing 
segment of the computer industry. This has 
created an ever growing need for programmer - 
system analysts, and many hobbiests or "hackers" 
see this as a means to earn money and take 
advantage of their programming skills. This 
paper will attempt to show the fledgling busi- 
ness programmer what he can expect from the 
small businessman, how to write proposals and 
run a business, and how to do small business 
system analysis. 

The serious hobbiest market appears to be 
finite. It is a market somewhat similar to the 
serious ham radio market in that it requires a 
great deal of technical proficiency and 
knowledge and previous experience in order to be 
successful. When the home computer first burst 
on the scene about two years ago, this was a 
natural market on which the new manufacturers 
concentrated. Of late, the realization has 
dawned on many of them that they will have to 
find another area in which to sell if they are 
to survive in a crowded marketplace. 

What Is The Market 

The Small Business Administration says 
that there are 13 million in the country. 
They produce 44% of the jobs and 36% of the GNP 
and form 97% of all U. S. business. The 
definition of a small business by this Small 
Business Administration is complicated and makes 
use of a number of criteria. These may be num- 
ber of employees and dollar volume of the 
business, and whether it is dominant in its 
field of operation. By that definition, a 
company such as American Motors could be 
considered to be a small business. The Table 
A-l presents some interesting figures, and 
Table A-2 is another set. You will note that 
there is little consistency in the numbers or 
definitions, and I would note that while the 
market is large according to all established 
surveys, I suspect it is not as accessible as 
most small manufacturers think. 

Characteristics of a Small Business 

Reams have been written about the 
definition of a small business and its 
characteristics. Its purpose is to produce 
on a timely consistent basis, at a profit, a 
service or product suitable for the intended 
use. It is usually managed and owned by a 
person who is a specialist in some particular 
aspect of a product or service. He or she is 
most usually a clever marketeer, or a 
superior technician or expert in his or her 
chosen field. 

Small business is a place for mature, 
intuitive judgment, not a place for the simple 
execution of a prescribed and defined routine. 
It usually exists to meet a specialized need 
which cannot be met by a larger company. 
Usually it fits into a chink or seam in the 
marketplace where size of the market is consid- 
ered not suitable for a larger company, or 
where the entry into a new market would pose an 
unacceptable threat to an already existing 
profitable large company market. For instance, 
in the computer business IBM is slow to 
introduce new technical developments because to 
do so prematurely would threaten their 
existing rental and lease base. In fact, the 
commercial smaTl business systems producers are 
in the same boat since they are eying some of 
the same markets that the small hobbiest manu- 
facturers are examining. However, the 
existing pricing level is much lower than what 
the commercial small business systems manu- 
facturers are accustomed to. In short, most 
small business companies, computer or non- 
computer are characterized by great flexibility 
and adaptability to rapid change. 

Characteristics of Small Business People 

Given the previous description of a small 
business, it should not be unexpected that a 
particular type of person will tend to 
gravitate towards this type of enterprise. 
They are more than usually independant and 
agressive, they are not very empathetic, and 
they mix aggressive behavior with a sense of 
dissatisfaction with their environment that 



BOX 1579, PALO ALTO CA 94302 

impels them to change things. Last, but not 
least, despite all of this motion and 
movement they are usually not to careful about 
detail. Their "span of control" is established 
by artful perception, intuition, and detailed 
experience. In a larger business this type of 
control is done by classical financial methods 
and budgets, or by formal analytical procedures. 
A person with this type of personality charac- 
teristics tends to change things rapidly and 
does not tend to think in algorithms. This can 
make a problem for a programmer who wants to 
work with small business people. 

On the other hand, the programmer - systems 
analyst tends to be the type of person who, 
while independent, is not usually aggressive. 
He does not have too much empathy and is 
usually satisfied with his environment. If he 
is to be a good programmer, he must have a 
high degree of attention to detail. If you com- 
pare these traits with the characteristics of 
the entrepreneur, you will find that the only 
thing they have in common is a lack of empathy 
and a failure to communicate. This may seem an 
over-simplification, but my many years of 
experience has shown me that there is this lack 
of similarity in outlook and objectives which 
cause many problems between the business man and 
the programmer. 

How Do You Then Merge The Two Interests 

First of all, the business man will 
appreciate your services more if you run them 
in a businesslike manner. I find reams of 
information that are written about programming 
skills, structured and nonstructured, and 
which cover the myraitf other factors that make 
up the tool kit of the proficient programmer 
and systems analyst. But nowhere do I find a 
rationally written instruction which tells anyone 
what they should reasonably charge. The basic 
rationale is important. How the individual 
applies it should suit his or her particular 
circumstances. What I shall present will equally 
suit the fledgling programmer - analyst as well 
as the small systems house or the full time two 
or three man partnership that programs for small 
business people. 

You start with 365 days a year and 
immediately you subtract 104 of those for the 
52 weekends. Most people calculate 10 paid 
holidays, or in any case the legal ones when 
nobody else is working, and add another 10 to 
14 days for vacation. This is a rough calculation 
but you are now down to about 220 saleable days 
per year. If one multiplies this number by the 
standard 8 hours per 'day, you get 1,760 hours. 
If, for instance, you were to assume a reasonable 
salary of $15,000 annually, that breaks down to 
an hourly charge of approximately $8.50 per hour. 
The catch is that that is at 100% efficiency and 

rarely does any systems house or consulting 
firm end up being able to charge for all of the 
existing hours. Some of this inefficiency comes 
from nonchargeable hours expended on a customer, 
or for administrative details such as sales, 
finance and/or research development for a 
saleable package. It would be my guess that 
most systems houses really do not function at 
much better than 50% to 60% efficiency. Seventy 
percent is a consummation devoutly to be 
wished. With this assumption, it is only 
prudent to say that the hourly charge should be 
doubled to approximately $17.00 an hour. That 
means that we are at a fairly high level in 
hourly charges even before we have added any 
classical overhead. 

Overhead can be anything from equipment 
rental, regular rent, automobile expense, 
telephone, paper and storage supplies. A 
little imagination should allow the incipient 
programmer to make a list of this overhead. 
Let us just use an additional $200 a month 
overhead, which we will distribute evenly 
between telephone and motor vehicle expense. 
If we take our mythical 1,760 hours and divide 
it by 12, we get 146 available hours per month. 
At 50% efficiency, that says we have about 
73 chargeable hours. As an aside, it would be 
the lucky new programmer that has more than 
half of his time chargeable when he is initially 
starting out. If we divide our miniscule $200 
a month overhead by 73 hours, that will give us 
approximately $2.75 additional charge; so, we 
are now hovering in the $20.00 an hour range 
for our services, and we have not even done 
anything yet. Most overheads, in reality, are 
not that low even for a part-time programmer, 
and certainly the efficiencies I am talking 
about are far from imaginary. So, it would 
seem for a full-time programmer, the charges 
have to be $20 -to $25 per hour, and the 
part-time programmer with no overhead perhaps 
could charge approximately $10 per hour. I 
would note that plumbers and electricians and 
the rest of service trades which are utilized 
by small business charge at least comparable 
prices. If there is any sin that the system 
analyst - programmer commits, it is under- 
charging for his time. The moral is, if you 
don't value your time, don't expect anybody 
else to. The second rule that I would for- 
mulate is that programming and systems analysis 
is a time-selling business very much like an 
accountant or a lawyer. If you are serious 
about entering this business, you should sit 
,down with a sheet of accounting paper and list 
all of your expenses and their monthly costs 
and go through an hourly calculation such as 
I have indicated in order to figure out what 
an adequate charge is to get a proper return on 
your invested time. 



BOX 1579, PALO ALTO CA 94302 

Why Consider A Small Business Micro? 

When you talk to your small business 
entrepreneur, you should use the same 
criteria as you think you would use for any 
other piece of equipment. He is going to look 
at this as: 

A. A labor saving device 

B. For an increase in productivity 

C. To provide better information 

There are two types of information he is 
likely to want. One type is an application 
that tends to have very heavy computational 
requirements and is analytical in nature, and 
has relatively light file or transaction volume. 
The second type of application is characterized 
by a heavy load of individual transactions, that 
are heavy file oriented, and with relatively 
easy computational requirements. These trans- 
action-oriented applications are the things that 
you hear all the time like payroll, accounts 
payable, accounts receivable, inventory, sales 
analysis, order entry, and P & L and General 
Ledger. The business man's rationale for using 
DP is that it is going to make things either 
easier, cheaper, faster or more efficient. Most 
systems analysts and programmers that I know 
lose sight of this. In this case, I am on the 
side of the businessman and against the systems 
analyst or programmer who falls in love with 
complexity for complexity's sake. The projected 
use should clearly effect economics in time, 
money, or effort, or it should improve produc- 
tivity. If it improves service or provides 
better information to make decisions, then so 
much the better. 

The way to start is simply to ask the small 
businessman what he is trying to do and to listen 
carefully. If you can analyze the flow of his 
product or service, you are at the beginning of 
your project. Most projects can be divided into 
four stages: 

Analysis and specification - 
Des i gn 
Codi ng 


If you look at this carefully, you will see that 
an analysis and design in any well executed 
project can take up to two-thirds of the available 

You've now listened to the businessman, you 
understand his product or services in some respect 
and he now looks at you expectantly with the idea 
that you are going to say something useful. Think 
about it. What are you going to say? What is more 
important, what are you going to do? 

I The first think that you should do is to say 
■that you will submit a written proposal. Emphasis 



on the written. It is a source of never 
ending amazement to me that many large or 
small companies that I have dealt with over 
the years that do not have written speci- 
fications and objectives for their computer 
installations. It is for this reason that 
McKinsey & Company at one point postulated 
that about 80% of the business computer in- 
stallations in the United States were 
failures. They defined a failure simply in 
the terms that the installation was not 
doing what it was planned to do. Installa- 
tions that were meant to provide total 
management information systems have in many 
cases degenerated into producing only pur- 
chase orders, payrolls, or inventory 
status. In many cases, these are high 
volume operations in large companies that 
have little or no relationship to what 
you are going to with a micro in a small 

I would approach the actual proposal 
in the following manner: 

1. I would have a conversation with the 
proprietor in which I would try to pin down 
that single thing which is impelling him to 
computerize. If it turns out to be "that 
everybody else is doing it", and that is 
the reason that he wishes to computerize 
something, I think I might quit right then 
and there. However, if he has a legitimate 
reason for wanting to computerize some part 
of his business, you should determine what 
are the priorities, and which things are 
most important. Try to concentrate on 

2. I would propose to him a study, for 
which he or she would pay , which would 
probably involve a minimum of 40 hours of 
your time both to prepare and document. 
This study would be paid for whether further 
work occurred on the system or not. As I 
pointed out earlier, probably 30% of the 
time that you should spend on a job is spent 
in specification, and is really the highest 
order of skill that you use. Coding is not 
the game in this case, and in fact if you 
start coding a point too early, you will 
probably create a system that is not 
responsive to your customer's need. 

The study should encompass the 
following areas: It should state explicitly 
at the beginning what the purposes of the 
system are to be, and additionally it 
should contain flow charts of the overall 
system and of the specific parts; also, 
some input specifications, which can be 
screen formats or anything else which is 
visible, as long as they are representative 
of the way the input really is to occur. 
There should be file descriptions and some 

BOX 1579, PALO ALTO CA 94302 

sort of output or report specifications. Again, 
they can be dummies. They do not need to be 
typed and they can be done on the standard report 
forms that most of the hardware companies produce. 
Again the simple fact is that they must be 
written down. There should then ' 2 a simple 
narrative as to how all this is going to be tied 
together. The proposal should be specified in 
this manner and the proposal presented to the 
customer. It should be paid for, and if the 
customer wishes to continue, he should sign-off 
at this point that this is what he wants. You 
should then provide him with an estimate for 
the hardware and the software to accomplish 
your proposal. It is obviously possible to 
include the hardware and software time and costs 
in the systems study, if that is what is 

A note on schedules. It is quite common to 
underestimate time. Most people can estimate 
costs better than they can estimate time; so 
that you must be very careful that you can 
accomplish the specified tasks in the proposal 
in the time and hours that you allow. It is 
not a bad idea to total the times required for 
the individual tasks in order to get an overall 
job time. By that I mean it is far better to 
specify a job piece by piece in terms of trying 
to determine how much coding there will be and 
then look at the total at the end, rather than 
it is to look at a job and say that is a 200 
hour job. Most of the time, those "gut feel" 
estimates are wrong, and it is compounded in 
many cases by the fact that you are probably 
undercharging as far as your hourly rate is 

At each point in the process when a 
particular system piece is completed, you should 
have specified ahead of time test data that will 
demonstrate that the programs are working as per 
specification. Note that I emphasize test data, 
and that I emphasize agreeing on this ahead of 
time as part of the programming proposal. The 
rationale behind this is that you cannot really 
be responsible for the whole of your customer's 
information base. You cannot spend enough time 
to verify the integrity of all his information, 
and you are depending on previously specifying 
what the system will accomplish. There is 
many times a large gap in what the specified 
system will accomplish and what the customer's 
hidden agenda is, is related to what he wants to 
accomplish. For the uneducated, naive first 
time user of computers, it has been made to look 
ridiculously easy by much of the advertising 
and much of the media. So, in fact, he has 
high expectations where he or she should not. 
If you want to do a good job for your customer, 
and stay in business, you have to make sure that 
everything is done on a realistic basis. For 
that reason, test data should be specified 
ahead of time, and when that test data is run, 

it should be considered that the job is 
compl ete . 

Changes are extra, and you should make 
this clear from the beginning. If the changes 
result from a customer change in specification, 
they should be paid for. You should estimate 
what they are going to cost before you under- 
take the changes. In order to be fair to the 
customer, he should know what his change is 
going to cost him before you start to work on 
it. Probably more disputes have been caused 
by the failure to do this and the failure to 
specify things in writing then anything else 
in the computer business. I would repeat that 
software problems with customers come from 
unrealistic specifications, and from an 
assumption that the customer knows what it 
takes to change something. He does not if he 
is not familiar with computers. He thinks 
that he can change it the same way you erase 
something with a pencil. He knows nothing of 
program logic, and he has no real realization 
of how long it takes to reformat a report or 
a file with 10,000 records. 

There are some additional problems of 
which the fledgling systems analyst programmer 
should be aware. 

He should be aware of his customer's 
company atmosphere. How good are the regular 
administrative procedures of the company? How 
much is in writing? Is there discipline, not 
in the sense of punishment, but in the sense 
of following established procedures. Are 
there controls on the activities that are 
performed in the company or is he winging it 
oh the basis of notes on 'backs "of envelopes? 
Is there any administrative backup to the 
proprietor? In some cases, this company 
atmosphere should give you a clue as to whether 
you can design a successful system of not. 


In this context, I am talking about the 
use of so-called hobby or amateur equipment 
as opposed to that which comes from the 
commercial market. The difference is sometimes 
more in the support and maintenance areas than 
it is in the actual hardware. However, the 
so-called hobby systems are plagued by poor 
support. The small businessman expects to 
have any type of equipment he buys supported 
by the people from whom he purchases it. In 
the micro-computer business, so far that has 
not been the case. Unless the local computer 
store is well organized and able to support 
the customer's equipment, the first electronic 
failure that he has is going to be a traumatic 
experience. If somebody tells him to bundle 
up his CPU and send it back to the factory, 
and it becomes clear to the entrepreneur that 



BOX 1579, PALO ALTO CA 94302 

he is not going to be able to access his 
inventory or his general ledger until this 
mysterious entity comes forth with a new or 
repaired board. He is going to be in a rage. 
Secondly, if he finds out that there is not a 
serviceman available to him who can come to his 
office to fix this, he will probably be 
displeased. His criteria is probably the 
electric office typewriter. If he has a failure, 
the serviceman comes and fixes it, and likewise 
with his adding machine, or his lathe, or his 
milling machine, or his stamping press. The 
micro manufacturers have not yet achieved this 

Another hardware limitation is lack of 
proper peripherals. In many cases, if you are 
going to use the system in even a very small 
business, there will be large requirement for 
printed reports. A proper high-speed dot-matrix 
printer is still an expensive item, and in most 
cases can come to at least 50% of the system. 
The CPU represents the cheapest part of the 
system, with some form of mass storage also 
being a necessity for most small businesses. 
Of late there seems to be a tendency for most 
people to include a double or triple floppy in 
the small business system. I would describe 
this as the minimum which can be tolerated, and 
I have severe doubts, even with very small 
businesses, whether they can last very long on 
a floppy based system. 

A lot of the wonderful stories of success 
about systems being applied to small businesses 
come from a technically oriented proprietor 
who bought the system, using his business as an 
excuse somewhat in the manner that larger 
companies buy corporate airplanes and depreciate 
them at the company's expense. He has done all 
the programming and is intimately involved with 
the success of the system. I don't consider 
this as a true business application of the 
hobby micro. There are very few companies in 
terms of the total number of small businesses 
available in the market that have that 
capability. It is much more likely that you 
will run into an entrepreneur with no previous 
computer training who simply wants a packaged 
program that can be made to work. 

For the foreseeable future we are not 
going to be CPU limited; we are going to be 
limited by the mechanical peripherals which are 
still purchased by the pound. I also have some 
considerable apprehension about hard disk 
systems. In many cases now they are being 
offered as an add-on to S-100 bus systems. The 
problem is that the disk operating system has 
not been designed as an integral part of both 
the drive and the CPU. This can make for some 
severe operating problems and I will touch on 
those later when I talk about systems soft- 

So, in general, we still have cheap CPU 
power; we con't have a good widely available, 
cheap, hard copy printer; and we don't have any 
standardized mass storage media. This 
situation should change when the electronic 
mass storage media become available. When 
either bubbles or CCD devices become widely 
available it will probably mean a big boost to 
the hobby-based system. There will be a big 
decrease in price because the fundamentally 
electronic devices are priced on a rapidly 
declining curve which is based on semi- 
conductor device yields. The electro-mechanical 
peripherals and storage devices are not subject 
to this kind of a learning curve. Therefore, 
their price does not decline as rapidly. 


There are probably only two fundamental 
differences between a micro processor based 
hobby system and a micro processor based commer- 
cial system. They are reliability and the 
presence of software utilities. Years back, 
when mini's first emerged, the situation was 
much the same as we find now with the micro's. 
We had a "gee-whiz wonderful" technical device 
that had very little in terms of peripherals 
support and nothing in terms of software 
utilities support. These utilities can be 
described as the tools which are necessary for 
the applications and systems programmer to use 
when he creates a system. It is probably not 
economically viable to write small business 
systems fromscratch without the aid of some 
powerful programming utilities. These are 
file managers, report generators, and data 
based systems. Interestingly enough, there 
appear to be a couple of companies in the 
hobby field that have worked out data base 
management systems for micros and I have seen 
some advertisements for some of these on a 
commercial basis. They are still not "poor 
man" systems, but they are cheaper than the 
commercial systems and they appear to have 
considerable capability. It would be my 
advice for any fledgling systems - analyst - 
business programmer to get some experience on 
an equivalent mini based system so that he 
understands what is available in terms of this 
type of software. Not only do the micro 
systems not have this kind of system software, 
but what they do have is nonstandard. This 
lack of standardization applies from every- 
thing from the version of basic that is used on 
up through the communications protocols. It 
is a "Tower of Babel" and the user and the 
systems analyst and programmer are suffering 
because of it. There really is not going to be 
any way to make a good efficient set of 
programming tools until the industry can get 
together and determine what standards it is 
willing to support. Up until that time they 
will be forced to look for commercial 



BOX 1579, PALO ALTO CA 94302 

packages and utilities. 

There seems to have developed a thriving 
business in package software. I have looked 
at much of this, and I am not impressed. Very 
little of it was written for the general case; 
most of it was written for one specific case or 
another, and it is not very adaptable or 
transportable because of the lack of standards 
and systems utilities that I mentioned. 




Total Businesses 13,343,406 

Corporations (1972) 1,823,335 

"Subchapter S" Corporations 

Partnerships 1,037,91 1 

Proprietorships 10,482,160 

Source: Preliminary Statistics of the Internal Revenue Service, 1973 

In short, with careful study you can be 
successful; but, it takes the use of an equal, 
or greater, dose of caution and business acumen 
as well as technical knowledge and programming 
skill . 


Over 20 million Small Bualnessea, 30,000 medium-size firms and the Fortune 1500 provide 
good* and services for our free antarpnaa economy. 17 million are sola proprietorships, 1 
nillim are Partnarahlpa fwtth an average of 3 partner* each) and almoat 2 million are Cor- 
poratlena. The following figure* ware compiled from I.R.S. Statistics of Income - 1968. 
"Nat Profit*" dees Loeaea). include "wages" to P r oc la im s (Schsd. O and Property Owners 
(Sched. Qt and are before federal income tax except for Incorporated businesses. About 
1/3 of U.S. Businesses showed Net Losses. Almost 700,000 new businesses are started 
each year, while 500,000 go out of business. 

Number of 


Gross Nat Profits Average 

Receipts minus losses Profit 

tmlUlviH) IbUUom) W?U«ra> 


Kentai Property Oners ™ S"o46 

Inventors b 2xplor*ra(Royaltlea) '50S 

TOTAL - Jig. ,Coaaerce,Servlcea, 

Construction, Transport «,F*raa u ,677 


Agriculture, forestry, flab 3,070 
Agricultural services 288 

Mining 68 

Contract Construction 840 

Manufacturing 397 

11. 800. 218 $128.600 S11.000 

Pood It kindered products 
Textile 4 apparel 
Limber a vood products 
Printing k publishing 

Fabric* ted aetai producta- 
Machlnery (non-electrical) 
Electrical Machinery 
Rubber , leather .atone 
Petroleua refining 
Primary Metal Industries 
Motor Vehicles 
Other Transportation equip. 
Mlac. Manufacturing 
Tvanaportatlon k Utilities 

Local Tranaportatloa 
Trucking a Warehoualng 
Electric, Gaa a Sanitary 
Hctall-Wholeaale trade 









Wholesale trade 

Building ■atari* la 

.General merchandise 

Food atoree 

Auto dealera 

Gasoline atatlona 

Apparel atoree 


Household appliances 

Eating places 

Drinking places 

Mlac. Stores 
Real Estate. Ine.. Finance 


Real eatate 














Hotels, Mote is. Trailer Parka a> Cai 

Personal aervlcea 
Laundry k dry cleaning 
Beauty shops 
Barber shops 

Business aervlcea 
Auto repair a, aervlcea 
Appliance e> other repair* 
Aauaeaent, recreation k theatrical 
Doctors, Dentists, Nureee k Health 
Legal aervlcea 
Educational services 
Engineering a Architects 
Accounting k Bookkeeping 
Other services 


47 , 277 



99, 039. 

' 43! 655 
52 , 348 

46 , 854 
34 , 992 

51 1 376 


20 , 826 
20 , 434 







"4" 247 











20 . 166 


59 , 500 











495 , 300 


























495 , 300 





















T55 - 

















" 2:900 


<o? 1. osusem w • o»c*co. u. «o*oi . -ww-c-f um or^ie* 



BOX 1579, PALO ALTO CA 94302 


Wm. J. Schenker, M.D. 

Medical Information Systems 

2086 Essenay Avenue 

Walnut Creek, CA, 9^596 

[k\5) 939-6295 


Earmarking capital for maintenance is 
one of the things which sets a business 
or commercial venture apart from a 
typical hobbyists activity. These cost 
factors can be dealt with in two ways, 
as a science and as an art. The former 
is the most visible, providing subject 
matter for textbooks, seminars, and 
college credit. Complex as it is, it is 
still based on the simple linear premise 
that 2+2 really does equal k, and such 
like. This makes it relatively easy to 
read and write about. 

On the other hand, what the science 
deftly avoids is a large object whose 
icy tip spells disaster in the deep for 
the unsuspecting businessman or 
professional. Etched in large letters 
on the submerged surface, hidden to 
ordinary view, is the warning, "Murphy's 
Law (and Cohen* s Corollary) Reigns Ever 
Supreme ! " 

It is this murky subject, the center 
of many a hallway conversation among 
insiders, and rarely discussed as the 
computer systems vendor plies his trade 
among the innocent, that will be 
emphasized in this paper. 

I. Introduction 

EXHIBIT A. There is a medical clinic 
which has had considerable experience 
with computers in the last decade. 
Three years ago-, when operational 
microcomputers were as rare and 
expensive as hen's teeth it had the good 
fortune to be offered on loan from a 
local government scientific organization 
a complete micro system of the highest 
caliber. Here was a chance to evaluate 
this new technology in a medical 
environment and for free. The clinic 
chief, wise in the ways of computer 
vagaries, when approached with this 
offer responded, "Sure we'd love to have 
use of such a system - but only if 
you'll pay the maintenance costs." 

have an electronics background." This 
challenge is not constrained by time 
factors nor by responsibility to an 
outsider such as your customer. Since 
time is money you can see how hobbyists 
can spend $10,000 in labor to repair a 
computer costing $1000. 

To the businessman or professional 
however, equipment breakdown means at 
best an added expense and at worst lost 
income and a blemish on the 
organization's market image. To 
appreciate the significance of this 
consider the following. When a computer 
system suddenly stops running, or "goes 
down" or "crashes" in the vernacular, 
the obvious cost is loss of service to 
the customer. What is probably in the 
long run a much more telling loss is the 
loss of data base integrity. This 
occurs when the transaction in process 
at the time of crash gets lost (or 
duplicated!), a record during update is 
lost, or a spurious record pointer 
change occurs. This last can result in 
possible loss of a massive number of 
records. The bottom line effect of all 
this on your business is loss of 
customer confidence. 

Let's look then at what must be done 
to maintain uninterrupted performance to 
your customer or client at the level the 
latter's accustomed to or contracted 
for. And some idea of what these 
actions will cost in the way of capital 
investment and added payroll. 

These questions are important even 
for a computer "application" or 
assignment which is only periodic in 
nature, such as getting a payroll out 
every two weeks. But they can loom 
large enough to become a pivotal factor 
in the organization's overall success if 
the computer's output needs to be close 
to continous. 

EXHIBIT B. A centralized medical 
boratory is planning a program which 
11 process up to 20,000 tests a day 
d return the results via computer link 

To tiie noooyist, so much in evidenc 

here at the Computer Faire today, system to outlying clinics the same day; any 

failure is a challenge which promises at breakdown in such a system for more than 

worst a broadening of knowledge, at its a few minutes could create a backlog 

best the ego rush of successful debug 
and repair of a non-working 
conglomeration of wires, chips and 
boards. "1 did it, and 1 don't even 


forcing delay of some results over to 
the next day. This would be considered 
intolerable in many cases by the doctors 
waiting for results. 

245 BOX 1 579, PALO ALTO CA 94302 

Thus the problem can be seen to 
warrant close study and inspired 
application. Indeed so much that 
journal articles and textbook chapters 
are dedicated to it. Now this 
literature tends to follow the formula 
of computer science publications in 
general. The orientation is extremely 
rational in tone and the reader is 
assumed of a likewise bent. The people 
in this field tend to have a heavy 
background in or orientation to 
mathematics. Thus you will find the 
literature also based heavily on a 
mathematical or statistical approach 
worthy of the physical sciences. At the 
heart of it all is the reasonable 
assumption that 2 + 2 really does = **, 
and such like. As a matter of fact the 
keyword here is reasonableness. It 
typifies the standard approach to 
maintenance strategy, the science of 
systems maintenance, and budgeting for 

Contrasting this approach is that 
ill-defined, scantily documented, and 
trivially regarded collection of 
anecdotal material and opinions best 
summed up as the art of systems 
maintenance and budgeting. It is this 
aspect of small systems or microcomputer 
technology that will be emphasized in 
this paper. 

II. Some Standard Recommendations 

A superficial appraisal of 
maintenance would focus on the obvious, 
the specs of the warranty and 
maintenance contract. The important 
facets of the latter include: location 
of service depot, minor parts or all 
parts covered?, charge for travel time?, 
preventative maintenance schedule 
included?, and is service agreement on 
an hourly ("on call") or contract 

However, experience dictates 
considering as well, the design and 
configuration of the system itself. 
Because these decisions made long before 
system purchase impact so heavily on 
subsequent problems, one must focus as 
well on these factors. Accordingly, 
note the following brief but 
representative list of points that an 
end-user would be advised to investigate 
prior to purchase. 

1. Buy from one vendor. A mixed-vendor 
potpourri will find inter-vendor finger 
pointing at the time of breakdown, each 
one accusing the other as the basic 

2. Vendor pedigree. The vendor should 
be big-name and well established, even 
tho the initial price is much higher. 
Avoid the fly-by-night and those without 
a track record. 

3. Vendor "burn-in".. Burn-in is the 
process of pushing the hardware close to 
its limits to see if it will stand up to 
prolonged temperature, vibration, 
humidity, dust, and electrical noise 
stresses. Investigate the details of 
this procedure by the vendor of your 

k. MTBF & MTTR statistics. MTBF is the 
mean time between failures, and MTTR is 
the mean time to repair such failures. 
Ask the vendor to show you his figures. 

5« Vendor warranty. Investigate 
carefully the vendor warranty details. 
Consider this an important step. 

6. Source of maintenance contract . Buy 
your maintenance contract preferably 
from the system vendor, or as a second 
choice, from one of the nationally-known 
service companies. 

7. Track record. Buy a system that's 
been out in the field with a long, 
reliable track record. 

8. Site of system. Buy a system larger 
and more powerful than necessary for the 
application at hand - to allow for 
subsequent expansion of the applications 
environment, thru the use of 
■multitasking" software. (This 
technique, allowing 2 or more jobs to be 
handled by one machine at times 
apparently simultaneously will be 
discussed in more detail later.) By 
planning ahead in this manner you can 
save on maintenance in the long run, by 
avoiding complex and unreliable 
retrofits, kluges and mismatches. A 
"larger than necessary" system here 
could mean a minicomputer, even tho the 
price is 3 to 10 times that of a micro 

9. "Diagnostics". Buy all the 
available software packages of this 
genre. They allow you to test out your 
hardware in a routine and thorough 
manner, often enabling you to anticipate 
system failures before they cause a 
total "crash". 

10. IC chips soldered in. Buy systems 
with the chips soldered, not socketed 



BOX 1579, PALO ALTO CA 94302 

in. The latter technique is a source of 
intermittent failures. 

11. Mass storage peripherals. First 
choice for small systems is the "floppy" 
or flexible disc, with digital cassette 
or 3M cartridge devices for backup. 

12. Memory. First choice is high 
density dynamic RAM, because of low 
power and chip count - therefore higher 

13. Printer speed. In general buy the 
fastest printer you can reasonably 
afford. (By the way, low speed in the 
big-name world means about 200 
characters per second or somewhat less 
than 200 lines per minute. Medium speed 
is about 600, and high speed goes up to 
an astronomical 30,000!) 

lk. Install an UPS. This means an 
uninterruptable power supply. It should 
be large enough to handle the power 
requirements of your entire computer 

15. Ambient temperature. Keep your 
computer system cool; use adequate 
convection in the form of fans inside 
the chassis, and air conditioning in the 

16. Nicotine. Avoid smoking in the 

computer room - the fumes are poison to 
magnetic disc and tape media. 

17. AC power lines and grounding. Use 
good filtering and solid grounding 
procedures, respectively. 

18. Modular vs all-in-one packaging. 
It's OK to buy the latter - it cuts down 
on troubleraaking interconnect cables, 
dust, meddling by unauthorized 
personnel, and generally makes for a 
neater appearance. 

19. Front panel. Your system should 
include this item. It's handy in 
troubleshooting (and incidentally helps 
in software debugging). 

20. TLC during infancy. There is 
another factor, not apparent to the 
newcomer in the field of EDP (electronic 
data procesing), which should be 
considered for its, impact on maintenance 
costs. Altho a system should operate in 
flawless fashion the first time it's 
powered up - that almost never happens. 
A computer system, like people, needs 
lots of tender loving care during 

infancy to ensure getting started on the 
right foot. Lacking this care, troubles 
will hound the system possibly to its 
grave. So until this stage is reached 
you can't consider visits by service 
personnel as part of maintenance costs 
and loss. When the system is completely 
debugged and performing to vendor specs 
for say a month, subsequent breakdowns 
are then properly in the category of 
"downtime", or failure. 

III. Observations on The Foregoing 
Plus Som e Maverick Recommendations of mv 

1. Single vendor systems. In the first 
place when you're dealing in small 
systems never buy from ANY vendor. Buy 
instead from your local retail computer 
shop. People you can talk to on a first 
name basis and whom you're likely to 
bump into at your neighborhood 
supermarket are much more likely to be 
responsive to your needs. In the second 
place if you buy what are called S-100 
products, you can count on a relatively 
high degree of inter-vendor 
compatability . More on that later. 

2. "Major league" vendors. There are 
no big name vendors whose primary 
business is small systems. Avoid the 
minicomputer firms and the big 
semiconductor manufacturers for whom 
micro systems are only a sideline - 
they're too big to care about you. 

Besides which, the "biggies" don't 
necessarily use the latest technology 
(because they feel somewhat immune to 
market pressures?). For example, the 
best support chip technology to day is 
Low Power Schottky (LS, for short). It 
is more reliable than its predecessor, 
TTL, because it produces less heat and 
electrical noise. 

support chips throughout. 

EXHIBIT E. Persci's floppy uses TTL 
support chips throughout. 

EXHIBIT F. NLS (a biggy in test and 
medical equipment) in their new model 
clinical lab unit uses RTL throughout. 
This technology, even older than TTL, 
was already vintage in 197^» when I 
first got into personal computers. 

3. Vendor burn-in. Temper the standard 
advice with the fact that the modern 
chip technology described just 
previously tends to be quite reliable 
once you're past the "infant mortality" 
stage. (This is chip failure within the 
first several hours of use.) If you do 



BOX 1579, PALO ALTO CA 94302 

feel you need it for your application 
environment don't rely on the micro 
vendors. They do very little of it. 
Instead have the local retail shop where 
you buy your system do it for you. 

k. Ah. the beauty of those neat MTBF 
and MTTR statistics! So crisp, so 
scientific, so neat, so precise. So 

At the outset the one thing to keep 
uppermost in your mind is that these 
MEAN figures apply only to Mr. MEAN 
End-user. He sits right in the middle 
of the bell curve of probability. Way 
off to one side of him is the fellow 
whose equipment never breaks down. But 
off to the other side is the fellow 
whose equipment fails before he even 
gets it out of the packing carton. Now 
no vendor will ever guarantee that 
you'll always fall between Mr. Mean and 
Mr. Superlucky - but many of their 
salesman will. But not in writing. 

To put it another way if you need to 
rely on 5 years between breakdowns you 
can't pick a system with a MTBF of 5 
years (even if such were available). 
You'd need a system with a MTBF of FIFTY 
years. To ignore this unpleasantry is 
courting a case of the "pre-demo blues" 
and full-blown operation of Murphy's 

EXHIBIT Fa. A large Bay Area service 
organization had planned for close to a 
year in 19?6 a demo for top echelon 
management of a complex multi-station, 
multi-tasked .system. As D-day 
approached everything began to fall into 
place, a tribute to the careful 
planning, competence and experience of 
the EDP (Electronic Data Processing) 
staff. By a week before the demo 
everything was pretty well sown up. 
Confidence was building for a successful 
performance on which would be pinned the 
hopes for the new budget. Confidence 
increased further in the last few days 
as everything continued to fall together 
right up to the eve of the big day. 
THEN the system crashed, too late to 
pick up the pieces in time for the 
visiting dignitaries, who witnessed a 
limping, anemic ghost of what was once a 
vibrant template for future glory. 

EXHIBITS Fb, c, and d. Three 
similarly painful experiences, on a 
smaller scale, with the first of my two 
systems in the past year. 

By the way, have you ever noticed 
something peculiar about luck, that 
phenomenon which is the" basic reference 
point for all statistical validation - 
except the statistician calls it 

"chance"? Anyway have you noticed how 
so much of life demonstrates that the 

lucky get luckier and the you 

know the rest? 

This may seem like rambling off the 
subject, but I'm reminded of how when I 
pull up to a toll gate pay station (or a 
supermarket checkout counter) I always 
look for the quickest line - and usually 
get the slowest. What bears mention is 
that altho I figured I must be a rare 
one, it's surprising how many others 
turn out to be members of the same 

And sort of inversely related to this 
is how a fellow intern back in the '50s 
made more money at, the end of every 
month than he got in hospital salary 
($125!) - by beating us all at poker for 
12 months straight. 

5. Vendor warranties. Forget it, in 
the micro field. They're worthless in 
terms of turnaround time (up to three 
months) - another reason to buy local. 
(See the following.) 

6. Maintenance contracts. The micro 
vendors don't offer them, and the 
national service companies don't have 
the necessary experience with the small 
systems. Again buy local, including 
your maintenance contract. 

7. Products with lone track record. 
Forget it. The entire market is only 
two years old. 

8. Buy a larger-than-necessary system? 
Don't, don't, don't. Buy the minimum you 
need for the application at hand. 
Discovered another application later? 
Buy another minimum system. For each 
added major application add another 
stand alone system, as close to 
identical as possible in hardware and 
software to your previous systems. 

Consider The Alternative. With 
today's labor costs in the form of 
programmer's wages, custom software 
implementation of the multitasking 
you'll need can cost 5 to 10 times the 
equivalent in hardware. 

More on Multitasking. Just to ensure 
proper understanding of this point a 
moment's digression into some details. 
This technique allowing two or more 
applications to be handled by one 
machine, at times apparently 
simultaneously, actually is a process 
akin to juggling 87 balls in the air at 
once - the computer's software 



BOX 1579, PALO ALTO CA 94302 

interweaving the multiple applications 
or "jobs" one between the other. The 
technique harks back to the days when 
the only hardware available were maxis 
and minis which cost so much per system 
that you were easily persuaded to 
squeeze out the maximum performance per 
piece of hardware bought. 

In fullblown versions used on the 
biggies it involves developing 
techniques for "queing" one job after 
another according to priorities, error 
checking of a complex nature, and 
complicated "rollback and recovery" of 
data when the system eventually crashes. 
It is responsible for a large software 
"overhead", i.e., software which is not 
earning you any money, while using up 
your computer's resources, both memory 
and processing speed or "throughput". 
Furthermore this is the kind of software 
that can take man years to develop and 
therefore costs plenty. It also helps 
to make the science of software 
development mysterious and their 
practicioners irreplaceable. Further, 
one would have expected by now something 
that complex and expensive would have 
made the large central computer systems 
"bullet proof" or impervious to error. 
It is certainly in part responsible for 
the somewhat unsavory reputation that 
computers have earned in non EDP circles 
over the past 20 years. 

EXHIBIT G. In July 1977 I applied 
for a Sears credit card. When it 
arrived it showed a purchase made in 
mid-June, 2 weeks before I ever saw the 
card. So I wrote to the special place 
you write to at Sears when there's any 
problem with your card. Well, they've 
been dunning me ever since, the monthly 
finance charge ever increasing. My next 
move is to cut the card up in small 
pieces, staple it all together with my 
latest bill and ship it off to Mr. 

EXHIBIT H. BART - spells Bay Area 
Rapid Transit, San Francisco's new train 
system. It also spells fiasco in 
connection with its originally designed 
computer control system. 

EXHIBIT I. Social Security. One of 
the computer scientists I work with 
knows of at least eight other SS card 
holders with his number. 

EXHIBIT J. Fill in your own 

9. Vendor diagnostics. There's hardly 
any available so have your local 
retailer write it for you. 

10. Soldered or socketed ICs. The need 



varies with local climatic and oi 
factors - let you local shop make this 

11. Mass storage peripherals. First 
choice is not floppies but MINI 
floppies, in particular North Star. For 
backup don't even consider digital 
cassette or 3M cartridge. Instead buy 
another North Star - it's in the same 
price bracket. If mini floppies don't 
meet your memory capacity needs bypass 
the full size floppy and go right to 
fixed discs. You'll thank me. 

How can 1 make such a recommendation 
in the face of the floppy's popularity? 
Well, the full-size floppy mechanicals 
are very tricky to align before you can 
get rock solid reliability. Then in 
about 6 months you may need it again. 
The mini floppies on the other hand have 
different physical dimensions (less 
inertia?) which make their tuneup easier 
to obtain and maintain. Where do I get 
this information? From end users, not 
from magazine articles. 

L2_. RAM memory pr eference. Spec your 

system to avoid dynamic RAM - it's too 
flaky with high speed peripherals using 
DMA (direct memory access), such as 
floppies and some graphics displays. 
Order instead FULLY static boards. Also 
avoid super high density boards such as 
6^K or 32K boards. Spec your system in 
increments of 16K. That way, if a chip 
goes bad on one board in a 6^K system 
you can still operate in a degraded mode 
with the other three. 

13. Printer speed. In general 1 
recommend the opposite: buy the slowest 
printer you can get by with. (What you 
can get by with may surprise you - as 
will be described later. ) Slower 
printers tend to stand up longer. 
Remember: "Speed kills." 

1**. Install an UPS. 

EXHIBIT K. Somebody accidentally 
trips on the line cord to the computer 
and pulls the plug. 

EXHIBIT L. Margaret, your secretary, 
is freezing because you won't turn up 
the thermostat to 80 degrees, so she 
quietly brings to work one day her 12 
Amp portable space heater and plugs it 
into the same 15 Amp circuit your 3 Amp 
computer is on. Everything's fine from 
9 to 930 AM with all transactions 
uneventfully processed and stored in 
RAM, at which time the program turns on 
your 3 Amp printer to get some hard copy 
and -- presto! -- the circuit breaker 



BOX 1579, PALO ALTO CA 94302 

pops and so does all your data from 9 to 
930 AM. Or worse yet, your disc 
operating software gets bombed, too. 

15. Keep cool. Yes. 

16. Avoid nicotine. Yes, it prevents 

computer cancer. 

17. Watch AC lines and ground well. 


18. All-in-one packaging of system. - 

INSTANT DEATH from a maintenance point 
of view. Go modular all the way. This 
point is so important that it will be 
covered in detail later. 

19. Front panels., Avoid them like the 

plague. The only one who might want one 
is your maintenance man, and he'll leave 
it in his toolbox on many jobs. 

20. TLC during infancy. Amen. Another 

reason to have service personnel close 
by, which means your local computer 
store, again. 

IV. Even Stranger Recommendations. 

1. "On Line" And "Real Time". 
Computer people bandy about two terms 
you should become familiar with, on line 
and real time. There are many 
definitions extant, but all you need 
remember is that some applications allow 
the computer to work in spurts, with 
long pauses for resuscitation in between 
- and at the other end of the scale some 
applications require the computer 
working 24 hours a day, 7 days a week 
with nary a skipped heart beat. Now the 
closer your application is to this 
continuous type of affair the more on 
line or real time it'll be considered. 

A piece of advice. If you're 
planning an application which could 
classify as pretty much on line or real 

2. "Non Stop". But suppose your 

3. The Nitty Gritty. Which brings us 
into the nitty gritty of my unorthodox 
approach to budgeting for small systems 
maintenance. My prescription for the 
typical business application which, 
altho not non-stop does have significant 
deadlines to meet, is simple. 

systems. Two computers, two sets of 
identical software, two sets of dual 
mini floppies, two backup storage 
devices (mini floppies again?), two 
keyboards, two video monitors, two 
printers, and lastly two complete sets 
of interconnect cables. (Ignore this 
last item and the whole deal is off. ) 

This strategy has three things going 
for it. The first is the ability to 
keep operating during a critical phase 
of activity when one computer crashes. 
True you lose the data that was in 
transaction and you can lose records, 
but with proper mass storage backup that 
barb can be dulled. Just flip the 

switch of the other system and transfer 
your work to it. (If you get your 
retail store to write some software and 
add some minimal hardware you can get a 
semi-automated transition from one 
system to the other.) You've spent 
twice as much in capital outlay in 
exchange for almost instant repair 
service, unobtainable any other way at 
any price. 

The second advantage has~~ta thr with 
that low speed printer I recommended 
earlier. By using your backup computer 
that is otherwise idle, to drive your 
printer you can overlap your application 
functions in time. For example you can 
be talking to one computer via its 
keyboard while the other is printing 
out your three hour report, but you 
could care les. In this manner you can 
usually do quite well with a 15 CPS 
(characters per second) printer where a 
60 CPS would be considered the bare 
minimum, or a 30 instead of a 120, etc. 
(This same strategy can be considered if 
you're trying to get by with a low speed 
mass storage peripheral, the audio 
cassette, which can be quite reliable if 
properly set up. ) 

application calls for nothing less than 

the extreme, a continous run? It's then 

logically enough, labeled non stop. 

More advice. If you're planning this 

kind of an application — STOP AND DON'T 


Unless your retail shop can configure 

your hardware and write enough software The third and most important 

to give you the micro systems equivalent advantage I save for last. It is 

-f what Tandem Computers, Inc. claims in key element in a novel concept of 

computer technology, geared to match the 
microcomputer's role in the increasingly 
popular EDP trend towards "distributed 
intelligence" thru "distributed 


their ads they can do with their large 
system. (If you're interested they're 
at 20605 Valley Green Drive, Cupertino, 
CA, 95014.) Their system is what is 
called "multiple redundant". 


processing". What these imposing 
250 BOX 1 579, PALO ALTO CA 94302 

phrases really mean is that instead of 
relying on one large computer to do all 
your work, you spread out some of this 
work by using a network of small 
computers scattered around the 

In effect you trust your eggs to more 
than one basket. This trend is having a 
salutory effect on the whole industry, 
making systems less vulnerable to "total 
crash". When the big one goes down 
people out in the boondocks can still do 
some work while waiting for the system 
to come back up again. It*s cheaper. 
It's also less complex and mysterious. 
(The enormity of the software problems 
associated with one computer doing 
everything was referred to earlier. ) 

Distributed Maintenance. I recommend 
we start doing the same thing, now, with 
the maintenance process. In a phrase we 
need what I like to call DISTRIBUTED 
MAINTENANCE. Spelled out this means 
that instead of relying on a central 
repair source (the vendor, or a national 
repair organisation) we get this service 
out into the field as close as possible 
geographically and timewise to the one 
who signs the bottom line, you the end 

EXHIBIT M. An excellent example of 
this kind of thing in actual practice is 
described by a resourceful Canadian, 
Jean Francois, in connection with a 
large minicomputer system he runs for 
the Ministry of State for Urban Affairs 
in Ottawa, published in DATAMATION, 
August 1977* This article should be 
must reading for you. 

But let's suppose you can't see your 
way clear to buying 2 of everything. 
Then with one system at your disposal 
your best bet would be to sign up with 
your local computer store for as close 
to complete coverage as you can afford 
(and the store can provide). The 
contract specs to check apart from the 
usual are: what is the guaranteed time 
to arrive, and what about nites, 
weekends, and holidays? Expect to pay 
from 10% to J0% of the system's purchase 
price annually, depending on whether you 
get minimum or "total" time coverage. 
With that kind of annual cost facing 
you, my proposal of 2 of everything 
seems less outlandish. 

Distributed Maintenance Protocol. 

Setting The Stage. The success of 
this strategy will depend on your local 
computer store for three things. 

1. A contract for repair of the 
defective equipment you track down with 
this method. It should specify the 
maximum "turnaround time" you think 
you're application can tolerate. One to 
three weeks will usually do for typical 
applications. (You'll save a bundle 
right there. ) 

2. Developing software, in the form 
of a short diagnostic package, that will 
tell you when in the course of your 
testing you've in fact bumped up against 
the troublemaker and have the system 
runnable again. This message will 
ordinarily be in the form of video 
screen prompts. 

3. Installation of an electronically 
simple yet very effective monitor that 
tells you if your power supply secondary 
voltage outputs (usually 3) are in good 
health. This will be in the form of 
little red pilot lights set into your 
(otherwise blank!) front panel. Or they 
can be installed inside the computer 

cabinet away from the high voltage end 
of your power supply, which should have 
a protective barrier placed over it by 
the shop personnel. This is so you 
can't monkey with it accidentally or on 
purpose. Using this last arrangement 
the lights should be in clear view on 
lifting off the computer cover. 

The importance of these monitor 
lights is this. If any of them are out 
it meansyou've lost one of the 
voltages; you should PROCEED NO FURTHER 
with subsequent tests, but get your 
maintenance man on the phone. 
Fortunately this will be a rare 
occurrence . 

First Phase. Swap modules or 
subsystems, one by one, FROM the known 
good TO the crashed system. Be sure the 
power is off both systems as you're 
making each swap. Off how long? Long 
enough for the capacitors to discharge = 
when the blower fans stop rotating. Do 
this until you find the trouble. Here's 
a good sequence to follow: interconnect 
cables, mass storage device, video 
monitor, keyboard, printer, and finally 
the computer itself. By now you'll have 
found out which module isthe problem. 
If it's any but the computer it's a job 
for your retail store. If it is the 
computer go on to the second phase. 



BOX 1579, PALO ALTO CA 94302 

Second Phase. Remove the cover from 
each of the computers. Start swapping 
boards, one by one, in the same manner 
you did with the modules previously 
(Remember, power off!). A good sequence 
to follow: memory boards, mass storage 
interface board, l/O board, and finally 
your CPU card or board. 

Simple Procedure. Remember, don't 
make this a complicated procedure - do 
it "by the numbers", preferably written 
down on a large cardboard placard placed 
on the wall near your system. Teach 
yourself the technique first, then you 
secretary, nurse, bookkeeper, office 
boy, or Girl Friday. A maximum of maybe 
l/2 hour using almost no technical 
expertise, is all you'll need to track 
down most troubles that can arise. You 
can then in a more leisurely fashion, 
send the defective piece of equipment to 
your retail shop where routine (and thus 
less costly) repair at the component and 
IC chip level can be performed. 

Equivalent Performance. To achieve 

the equivalent performance in terms of 
ultra short downtime and repair of 
defects would require keeping a full 

time computer tech on your premises (and 
payroll), AND he would have to have 
available a full set of replacement 
parts to achieve his goal. That salary 
in today's market is $15,000 and up a 
year. Then add the equivalent of "2 of 
everything" anyhow. 

Cheaper in The Long Run. Distributed 
maintenance, made possible by modern 
technology, will move computer science 
considerably closer to the kind of 
performance you the businessman or 
professional, expected to get in the 
first place. This is the kind of 
performance which spells business or 
professional success, which translates 
to more income. To boot, distributed 
maintenance is cheaper in the long run 
than any other method. 

No Fancy Test Equipment. While on 
the subject of cheaper let's cover a 
related point. Don't be talked into 
buying fancy and expensive products that 
enable you to troubleshoot like the pros 
do it. That's NOT what distributed 
maintenance is all about — . it's about 
non-electronic people using their time 
to get on with their own profession. So 
don't buy a scope, a logic probe, or a 
logic analyzer. And don't get into 
swapping IC chips either - you could 
blow a good one after a bad one that 


k.m Maintenance in a small town. With 
all this reference to relying on your 
local retail computer store, what if 
you're in a town with no such source 
available? Well in that case what might 
otherwise appear to you as a luxury, 
distributed maintenance, becomes a stark 

One Proviso. Even with 2 of 
everything don't even consider a 
near-non stop application if you're 
located in places like Last Chance, 
Kansas or Winnemucca, Nevada. There 
you'll need THREE of everything. Also 
you'd better consider making a special 
(and costly) contract with the nearest 
retail store with provisions for (a) the 
technical personnel remaining on site 
until the system's thru its infancy and 
TLC period, and (b^ paying them only 1/2 
the total purchase price on delivery, 
the other half after certain clear-cut, 
mutually agreed upon tests can be passed 
by the system's operation. If you can't 
get this kind of arrangement consider 
(a) foregoing the pleasures of rural 
life and moving to or near an -ugly big 

city or (b) running your business as 
before, in the manual mode - and buying 
a cheap computer for use at home and 
calling it a hobby. 

5 « The weirdest recommendation of all. 
This one's saved for last, since it's so 
obviously beyond the pale in our culture 
where rationality is considered the 
final criterion of any scientific 

Right at tile outset of systems 
planning, before you even look at your 
application needs, step back and ask 
yourself this question, and then answer 
it as honestly as you can. "Am I (a) 
consistently lucky in business or 
technical ventures, (b) lucky as often 
as unlucky, or (c) consistently 

If the answer is (a) then much of my 
ramblings can be ignored. If it's (b) 
it'll pay you to reflect on them. If 
it's (c) you're in the same boat I'm in, 
and to ignore my warnings augers well to 
bring you deep grief and near insanity 
in the form of slipped schedules and 
broken promises. (Take heart tho in the 
maxim, "Unlucky in technology, lucky in 
love! " ) 

V. Summary 

This paper has emphasized the more 
quirksome aspects of small systems 
maintenance problems. The major points 
made were: 

BOX 1579. PALO ALTO CA 94302 

1. Cost-effective maintenance decisions 
depend heavily for their success on 
making the correct choices long before 
the system goes down. As a matter of 
fact, it is at the time of original 
spec'ing out the system to be purchased 
that most of the die is cast. 

2. The single most important aspect of 
the system specs includes 

a. depending on purchase of 2 of 
every piece of equipment, thus allowing 

b. the full exercise of the 
distributed maintenance concept. 

Spelled out for small systems it 
means doing in-house swapping of 
interconnects, modules, and boards, 
backed up by a firm contract with a 
local reputable computer retail store 
for actual repair of the faulty 

VI. How to Evaluate This Paper 

After reading this far if you're a 
businessman or professional concerned 
about how he spends his dollars, you 
should be asking the question, "Sure 

this presentation is witty and 
provocative, but does he know what he's 
talking about?** And if you ask this 
question you've got to ask the next 
question, "Who can I ask to get the 
answer to my first question?" 

The Experts. So you're left to turn to 
the "experts" and the "authorities" in 
the field. Several things to remember 
about these fellows hov?ver, 

1. The field is so new there hasn't 
really yet developed a large cadre of 
knowledgeable neople. 

2. Then there are those experts whose 
expertize is in the mini and maxi 
computer field. Therefore they're 
likely to give you the standard list of 
recommendations that are valid in their 
world. Ipso facto, most of my 
unconventional recommendations will be 
"thumbs down" for them. 

3. Of those real experts we're looking 
for there are 2 general categories: 

a. The ones who've decided to 
capitalize on their know-how - they 
become vendors and computer store 
proprietors (I'm fortunately acquainted 
with some real honest ones, but you've 
got to be knowledgeable in the first 
place to recognized this honesty. ) 

b. The other kind who are hard to 
find. They're usually not in the market 
place selling their know-how but busy 

in their labs having fun. You won't 
find them in the yellow pages. 

The 87-foot monster. But wait, 
there's more yet. Filling this void 
then is the 87-foot monster I've been 
making the butt of my argument 
throughout this paper, the vendor 
salesmen. He's dangerous for you 
computer system's welfare, because - 

1. He's more accessible - you'll run 
into him everywhere: at the trade shows, 
on TV and Radio commercials, and in the 
slick business magazines. 

2. He's got less scruples. 

3. Mathematical and statistical "facts" 
are great tools in his hands to blur 
your vision of the nitty gritty you 
should be appraising instead. 

The Saving Grace. With all the caveats 
just mentioned where then can you turn? 
The answer is actually pretty straight 
forward. Go to your local retail store 
and ask for a list of the purchasers of 
10 or 12 complete business or 
professional systems they've installed 
prior to 6 months ago, and are presently 
serving on a full maintenance basis. If 
they haven't got that many to show you 

chances are you shouldn't be doing 
business with them - they haven't got 
the experience. Customers more recent 
than 6 months don't bother with - that 
doesn't give Murphy's Law long enough to 
rear its ugly head. 

Then make up a short polite note 
saying you'd like to talk by phone to 
them for 3-5 minutes the following week 
to inquire about maintenance experience 
and costs with their system. When 
you're done making these dozen calls 
you'll know whether I know what I'm 
talking about. 

VI. Cohen's Corollary 

I've enjoyed writing this paper and hope 
you have enjoyed hearing or reading it. 
But the alert among you may have noted 
there's still one item mentioned in my 
introduction that has yet to be laid to 
rest, namely the enunciation of Cohen's 
Corollary to Murphy's Law. You can't 
have heard of it before, because this is 
its first public proclamation. It goes 
like this. "When you'Ve taken that very 
last precaution possible to prevent 
Murphy's Law from operating in your 
applications environment — THAT'S when 
it probably will." 



BOX 1579, PALO ALTO CA 94302 


Gene Murrow, President 
Computer Power & Light, 12321 Ventura Blvd., Studio City, CA 91604 


Computer Power & Light, Inc. has been 
installing micro- computer systems in 
businesses since Fall of '76. This 
presentation describes four aspects of 
our experiences: a short description of 
the hardware we use; some actual case 
histories of customers for whom we've 
provided business systems; some of the 
jobs we've turned down (which is almost 
as illuminating as some of the jobs 
we've accepted); and fourth, some of 
what we consider to be the important, 
but often overlooked aspects of 
m i c r o - c om p u t i n g in business 


It seems to me that in the heady 
atmosphere of the growth of this new 
industry, a lot of claims are being 
made that are somewhat fantastic. We 
hear phrases like "what you can do with 
these are only limited by your 
imagination", at worst, to the no less 
extravagant, but potentially deceptive, 
claims that you can "do your payroll" 
with a $600 ..00 computer. Computer Power 
and Light, as a company, has had to 
face the music. We have had the 
customers coming in, magazines rolled 
up in their hands saying, "I want a 
computer that will do all these 
wonderful things." We've had to 
educate this customer, sell him a 
machine when appropriate, and then make 
sure that it worked and kept working. 
That's the basis of the experience that 
I'm going to relate. 


The hardware we work with is our own 
Compal-80 microcomputer. Compal is a 
contraction of Computer Power & Light. 
The machine is an 8080 based micro. It 
has a serial interface for devices like 
Xerox Diablo printers, DECwriters, 
modems, and others. It has a video 
display of 16 lines by 64 characters, 
an operating system on ROM, an anywhere 
from 32 to 56K of memory. We also 
incorporate the Micropolis dual mini- 

floppy disk drive. This is a high 
density drive; each diskette holds 315K 
bytes or characters of information, 
which is approximately 150 single 
spaced typewritten pages. We use 
either the Diablo daisywheel printer 
for our word processing applications or 
the Texas Instruments 810 printer for 
business applications. This is not a 
mini-computer, it is a micro- computer, 
and as such has all the price 
advantages and speed disadvantages that 
micro- computers have. A 56K system 
with the Diablo printer costs $8,605. 
A similar system with the TI 810 matrix 
printer costs $7,300. Of course 
there's lots of good hardware to be 
found. It's the software and the 
support and the other things that get 

Word Processing 

Now for some of the applications with 
which we have been successful and that 
you might consider as possibilities for 
a micro- computer system in business. 
The first application is word 
processing. It seemed like a natural 
to us. There's no number crunching 
really so the speed of some of the 
single chip micro- processors wouldn't 
be a factor. Our word processor, which 
is written to our own specifications, 
is written in the machine language of 
the 8080 and is consequently very fast. 
It is used to create business letters, 
assemble long documents from boiler 
plate mater ial , retype multiple 
revisions automatically, send the same 
"original" letter to each of hundreds 
of names on a mailing list, and index 
archived documents. It has essentially 
three functions: an edit function, a 
print function and a storage and 
retrieval function. In the edit mode 
you can enter text in a normal way on 
the typewriter, scroll text on the 
screen, search and replace a word or 
phrase throughout a text, "cut and 
paste" pieces of text (re-arrange 
blocks of text), bring in boiler plate 
paragraphs that are stored in a 
diskette library to assemble long 



BOX 1579, PALO ALTO CA 94302 

documents from boiler plate material, 
retype multiple revisions 
automatically, send the same "original" 
letter to each of hundreds of names on 
a mailing list, and index archived 
documents. It has essentially three 
functions: an edit function, a print 
function and a storage and retrieval 
function. In the edit mode you can 
enter text in a normal way on the 
typewriter, scroll text on the screen, 
search and replace a word or phrase 
throughout a text, "cut and paste" 
pieces of text (re-arrange blocks of 
text) , bring in boiler plate paragraphs 
that are stored in a diskette library 
to assemble long documents and things 
like that. In the print mode, you can 
set up your margins, your spacing, 
whether or not you want the right 
margin justified, you can have four or 
five different margin formats going at 
the same time, you can have variable 
character spacing, variable page 
lengths, variable line lengths and you 
can mix these up anyway you like, 
throughout a document. Finally, in the 
storage and retrieval mode, you can 
take a document, whether it's a full 
report or a paragraph or whatever, and 
store it under a name that's up to ten 
characters long, and then retrieve it 
just by typing that name on the 

So that's what we're doing with word 
processing. We had one eye on the 
Vydek and Lexitron systems when we 
designed it and the other eye on the 
pricetag. Our word processing systems 
start at about $6,0 r 0.00, which gives 
all of us micro- computer people a 
tremendous advantage over the so-called 
blind or non-video systems, such as the 
IBM mag-card systems or the Xerox 800 
system. And, we're about half the 
price of the large video oriented 
systems. Most importantly, the WORDPAL 
(as we call it) is just a program which 
runs on the general purpose micro. 

Secretarial Services. Who's using 

WORDPAL? Well, out in Van Nuys, 
there's a secretarial service called P 
& S Services. This was the first 
computer system we've ever sold to a 
woman. Harriet Wright became very 
knowledgeable about computers and 
looked all around before coming in. She 
told us the plight of secretarial 
services. They have very demanding 
customers who want the document today. 
If they wanted it tomorrow, they would 
bring it in tomorrow; that old story. 
The work has to be done rapidly and 
accurately. If the service isn't 

accurate the customer would type 
documents himself. Oftentimes they'll 
work all day on a document only to have 
the guy come back with the thing with 
500 different modifications that he 
wants. Or he wants all of the dates 
changed or something like that. So they 
need a flexibility for multiple 
revisions. Also, P & S needs to 
maintain steady customers who expect a 
consistent quality from her and 
consistent number of formats for the 
type of document they're going to be 
getting from her, whether they're doing 
speeches or manuscripts for a 
particular publisher, or whatever. So 
she was in the market for a word 
processor, and being a small company, 
like many of these secretarial services 
are, she wanted something she could 
afford and that had all of these 
capabilities. And, that's why I think 
she came to us. Another interesting 
aspect of the secretarial service 
business, and all you consultants 
listen hard, is that they all have 
delusions of grandeur, good delusions 
of grandeur. They would all like 
eventually to be business services, not 
just secretarial services. They'd like 
to offer bookkeeping help and record 
keeping and consequently they are very 
interested in the data processing 
capabilities of the micro- computers. 
So, the final clincher for P & S 
Services for us was that they knew that 
they could, by just changing the 
diskette, go from word processing 
capability to data processing 
capability. In fact, they are now 
starting to use our client accounting 
package to keep the books of the many 
customers whom they also serve as a 
secretarial service. 

Lawyers . The other big market for word 
processing, especially in the micro 
field, is lawyers. The main thing 
lawyers do, as we all know, is collate 
things they've used before. There's 
all this boiler plate that's sitting 
out there that they just like to pop in 
to a document, one section right after 
the other, and assemble up a set of 
interrogatories or a contract or a 
will. We offer them the capability of 
storing all that boiler plate on our 
disk system. Once the appropriate text 
is booted in off the disk, into the 
computer, they can go through and tell 
the computer that wherever you see the 
name Gene Murrow, replace that with 
Adam Osborne because I'm on Adam 
Osborne's case today. And that's 
precisely what they do. 



BOX 1579, PALO ALTO CA 94302 

We hooked up with one lawyer, in 
particular, who's been very influential 
to us and has helped us develop a real 
first class legal package. His name is 
Tom Lambert. He specializes in 
personal injury cases arising from 
aircraft accidents. He's a very smart 
guy; he and two of his colleagues have 
engineering degrees and have put in 
their years at Lockheed before entering 
law practice. When they go after the 
biggies like Cessna or Bell Helicopters 
(whom they specialize in), they can 
bring to bear a lot of their 
engineering capability, because they 
can evaluate rollover rates and all the 
bad things that happen to helicopters, 
for example. But the problem was that 
there was only three of them; when 
they'd sue Bell Helicopters, Bell would 
come in with their corporate law staff 
of 407 typists, 3^3 clerks and the 
rest. Tom shared his typist with the 
other three man office down the hall. 
He saw our system as a tremendous 
equalizer . Now when he takes on Bell 
Helicopters he can spend his time 
worrying about the engineering data and 
his overall plans while the machine 
automatically runs off fat books of 
interrogatories, which is what lawyers 
do to make life nasty for the 
opposition. The interrogatories often, 
in an aircraft case, run three and four 
hundred pages per plaintiff. Often it's 
a lot of boiler plate and cutting and 
pasting. He used to take three weeks to 
do each plaintiff's set of 
Interrogatories , just in the di soever y 
process (which is the opening salvos). 
But now he's got it down to about 3 
days per book, from three weeks, a 
factor of seven. Plus, he's free now 
to concentrate on the engineering data. 
Once again, I think the clincher was 
that he saw the Data Processing 
capabilities. So, when he's done 
preparing the interrogatories and is 
ready to go to the brief, or whatever 
it is, he can put in the Data 
Processing disc. He has written with 
his colleagues several engineering 
analysis programs that will model what 
happens when a 747 hits another one 
broadside or what happens to a 
helicopter when one of the rotors 
begins to loosen. These have become 
critical to his profession as he's able 
to put out graphical displays of this 
data that a jury can understand. So 
he's got the equalizer between him and 
the large law firms plus he's enhancing 
the impact of his own evidence, using a 
micro- computer, right where it counts 
- - in front of the jury. 

Academics . The third application for 
our micro-systems and word processing 
has been among academics: professors, 
departments and universities. They 
have a tremendous need for storage of 
yearly updated documents such as lists 
of required courses or lists of who's 
on leave this year. Plus, the 
department secretaries are under 
tremendous pressure to relieve the 
pressure that the professors are under 
in the "publish or perish" syndrome. 
They've got to get a few articles out 
to the journals every year and they 
depend on the department secretaries to 
do that typing. We hooked up with a 
fellow by the name of Dr. William 
Oldendorf at the V.A. Hospital in 
Westwood. He told us all about lab 
reports and articles to journals and 
things like that and once again he 
purchased a micro- computer based word 
processor so he could do those reports 
and submit those articles to the 
journals. When the editor comes back 
with 112 suggested modifications, which 
a typist can't do, he can sit down and 
waltz through the manuscript on the 
video screen, making the corrections. 
He then hands the diskette over to the 
typing department or the typist who 
prints it out. It's been a tremendous 
help for him. 

So that's word processing in three 
actual applications where people are 
saving money and are enhancing their 
businesses right now. They're not 
playing Star Trek with -these machines 
at all. 


The second major area, of course, is 
accounting. We wrote a client 
accounting package because two or three 
accountants walked into our store and" 
insisted that they could tell us 
everything we needed to know. So we 
worked with these three accountants and 
came up with a General Ledger or 
Bookkeeping package that allows you to 
set up your own chart of accounts and 
specify what format you want your 
reports in. , You then enter your 
transactions on a daily basis or weekly 
basis or hourly basis. And, at the end 
of whatever your reporting period is, 
whether it's a day, a month, a year or 
ten years... it'll go through the usual 
procedure. It does a trial balance in 
about 70 seconds. And, if everything 
eventually balances, (which you hope it 
does) out come an income statement, a 
cash receipts journal, check register 



BOX 1579, PALO ALTO CA 94302 

journal, five other regular journals, a 
complete General Ledger, a balance 
sheet, any schedules that come off the 
balance sheet and any subsidiary 
ledgers. All told it takes about an 
hour to run out all these reports. 

One of the accountants that came in was 
a woman by the name of Audrey Roche. 
Mrs. Roche is a professor of 
accounting, in fact head of the 
department, at Santa Monica College. 
She also runs her own business which 
provides the accounting services to her 
own clients. She was the one that 
helped us the most in developing this 
package and she's running it right now. 
She has a small service bureau which 
does people's books. She wouldn't have 
been able to compete doing it manually 
because a lot of the service bureaus 
now are automated . So , she wanted to 
automate and still have her own 
business. And, that's why the micro 
became important. A nice side benefit 
of all of this is that she sees the 
tremendous movement toward this 
solution. Santa Monica College is going 
to be among the first colleges in the 
area to offer a course specifically 
aimed at small computerized accounting 
systems. We're looking forward to 
providing several accounting systems 
for the college so that they can do 

Real Estate Investing 

Another interesting application that we 
found and that we're able to provide 
micro-computer support for was the real 
estate syndicators. These guys abound 
in Los Angeles, where even the Arabs 
are buying up property at a horrendous 
rate. What a real estate syndicator 
does is he gets a group of people 
together and says "lets all pool our 
money and buy this building and make 
lots of money." The trick is to 
convince everyone that it's a good deal 
and not to drop out. So, what they 
need, of course, is a forecast, a 
spreadsheet, that tells what's going to 
happen to an income property over the 

We sat down with two real estate 
brokers/ syndicators. It was one of the 
most interesting weeks we spent at the 
store, and came up with a package that 
does the following: 

You enter the purchase price of the 
property, the down payment, the 
depreciable basis, rate and term, (if 

you don't know it you enter some data 
from the Tax Assessor) . Then you enter 
a whole bunch of financing instruments, 
which is real estate jargon for loans. 
The data you enter is the amount of the 
loan, the term of the loan, the 
interest rate. That alone would have 
been easy except these guys are always 
talking about refinancing, balloon 
payments, variable interest rates. For 
example, you can start with three loans 
and in month 27 let's say, pay off two 
of the loans with a third loan at a 
different interest rate, pull some cash 
out of the deal at that point to pay 
for the air conditioning system that 
you're going to put in at that point, 
which you're going to depreciate for 
ten years to get a tax write off in 
year five... So, those are the deals we 
had to sort through and which caused 
our programmer to lose most of her 

After you do the loans you then have to 
do an income package. You tell the 
machine what the gross income of the 
building is or what the income per unit 
is and it prints out a nice list of who 
the current tenants are, what the rate 
per square footage is on the building 
as a whole, who's way under, who's way 
over, and what the vacancy rate is. 
Then you put in an expenses package. 
There are 20 Categories, everything 
from taxes and maintenance on down to 
the gardener and the swimming pool 
maintenance guy. What the program does 
after it chews up all this data is a 
year by year forecast or a forecast for 
a particular year that gives you the 
bottom line — the return on investment. 
That's calculated in several ways: the 
income versus the cash you put in, or 
the build-up in equity, or income vs. 
equity, and so forth. 

One of the biggest features of it is 
that a syndicator can put in the tax 
bracket of any investor before this 
thing flies off and it'll tell him the 
implications on his personal taxes. 
That's very important for doctors and 
others who often end up with too much 
money at the end of the year . They want 
to know what the implications are on 
their own taxes if they invest in this 
deal. The program also computes 
post-tax and pre-tax spendable and post 
tax and pre-tax income as far as the 
IRS is concerned. 

Another nice thing that it does is 
allow you, after printing out any 
forecast, to manipulate any single 



BOX 1579, PALO ALTO CA 94302 

variable to change the picture. 
Suppose we fired the gardener, how 
would that impact our income in the 
year five? 

The syndicator who was the first one to 
take delivery of the system about five 
months ago was Sheldon Allman in North 
Hollywood, California. He is a former 
stand-up comic. I don't know what that 
says about real estate. But, one of 
the things that interested him about 
the machine was sort of the inverse of 
what I was talking about before. He 
finally got sold on buying the machine 
because of its word processing 
capabilities. When he's all done 
running out all of these forecasts on 
different income properties he then 
turns to the word processor to write a 
very nice cover letter and cover 
description that goes "personally 
addressed" to each of his prospective 
clients as "this is a really good deal 
for you". So he uses both capabilities 
of the micro-computer . 


The final thing I'd like to mention in 
applications is in retail systems. 
This is a current project of Computer 
Power & Light. We've done the General 
Ledger, we have already done a payroll 
package and we've done some mailing 
list packages. We have finished up a 
sales anaylsis and marketing survey 
package and we're working on order 
entry, invoicing, aging of receivables, 
and payables package. 

The person who got us into that whole 
business is Harry Margulies. Harry is 
the owner of Beverly Stereo and 
Electronics, which is one of the 
venerable firms in Los Angeles. They 
were in it very early on when stereo 
and hi-fi was much the way micro- 
computers were a year ago — very much a 
new thing. Now he has a large number of 
employees with varying pay modes plus a 
tremendous inventory problem 
controlling the large number of items 
that a stereo store handles and all 
kinds of other things where a micro- 
computer could save a lot of time and 
money. Harry, right now, is running 
our payroll system, which handles the 
payroll for all of his employees 
weekly. He also uses our mailing 
system, which is part of the word 
processing packaging, to do direct mail 

Just to give you a brief description of 
the payroll program... it maintains a 

data base with employees' names and 
whether they're on an hourly rate or 
salary or commission or get a 
guarantee. It has all of the tax 
algorithms in it, not the tax table. We 
worked back from the tax tables and 
figured out what the algorithms were. 
They match in every case. It 
automatically will select the list of 
employees at pay day so you can run 
through and just tell whether he was 
there that day or that week or whatever 
and it prints the checks. It does the 
quarterly reports, the year end reports 
and it updates the data base. 

So that's some of the applications, 
some of the things we've done. It 
maybe fleshes out some of the words 
like "limited only by your imagination" 
or "this thing will balance your 
checkbook" or "play Star Trek" or 
whatever. I hope I've given you some 
feel for the kinds of things we're 


Now I'd like to talk briefly about the 
limitations. I'm going to tell you the 
jobs we turned down. And that I think 
any self- respecting micro- computer 
dealer should turn down. 

They fall into two categories. One is 
that the job is too complex. The 
software is just too complex and any 
decently written package would so swamp 
the cost of the micro-computer that you 
might as well, since the cost of the 
machine is in the noise anyway, buy a 
mini-computer, a faster and bigger 
machine. That may sound heretical but 
that's it folks. 

The second category is that in some 
cases micro- computers are just too 
slow, or the storage is too limited. I 
mean we've had people coming in with a 
check made out for $2,000.00 who 
actually needed 128 megabyte disk 

Here are some of the ones we've turned 
down. Medical Group Accounting. A 
doctor would come in and say "Hi, I 
work with five other doctors in my 
building and I'd like you to. ..I've 
heard about these micro-computers and 
I'd like you to do a system that'll 
compute all of the Medi-Cal and 
Medicare payments, all the insurance 
reports, keep all of my appointments, 
age my receivables, keep all of the 
patient records, and... " you know, 37 



BOX 1579, PALO ALTO CA 94302 

other assorted jobs. I think the 
medical thing is going to be a tough 
one to crack and we're going to have to 
work on it over the next few years. 
But right now I don't see any 
micro-computer really providing a cost 
effective solution to the entire 
medical problem. 

The second one we turned down was 
scheduling classes for a school. I 
don't know if you're familiar with what 
it takes to do that. You might have a 
thousand students and 300 teachers and 
275 classrooms and 6 periods per day 
and 1200 different courses. We had 
several people come in, from local 
private schools especially, who said 
"I'd love to have one of these 
micro-computers and this is what I want 
it to do. Build a master schedule, 
sign up the students into various 
blocks once the master schedule has 
been built, then print out rosters of 
every class so the teacher on the first 
day of class knows who's in his class. 
Plus, the room assignments for the 
students. And then every marking 
period I want all the grade reporting 
with class by class averages and 
department averages and all of that." 

That kind of an application will hurt 
the industry if we attempt it right 
now. Because people are going to say 
it doesn't work. I'm convinced that it 
won't work right now. 

The third area that we've avoided, that 
I feel represents a limitation in the 
micro- computer in a business area that 
we're all interested in, is 
point-of-sale terminals and other 
on-line or real time applications. It 
brought down a company by the name of 
Singer and I don't expect many of the 
micro-computer companies to attack that 
one right away. The micro-computer is 
ideal for it but there are some other 
problems. I don't know how I'd like to 
handle a phone call from a customer who 
calls up and says there are eight 
people waiting in line to buy something 
and the machine isn't working right. I 
think there are so many other factors 
in this that we had better tread 
carefully before we all... with new 
equipment .. .put in on-line real time 
systems. Even though I know the 
equipment can do it , we have to solve 
some other problems. We have to mature 
a little bit as an industry before we 
tackle that one I think. 

Other Considerations 

Finally, the last topic— some of the 
factors in our experience that, as a 
retail store, meeting businessmen and 
being out there amongst them, we've 
discovered that are very, very 
important and that aren't immediately 
obvious to the person who's delighting 
over the instruction set of an 8085 
chip or something. 

Service . One of them is service. Do we 
emulate IBM? Do we say to our 
customers "if anything goes wrong pick 
up the phone and we'll be there in an 
hour". To some extent, we have to do 
that. If we are telling a businessman 
that he's going to place his life 
savings, the hard work of himself and 
his wife, and the success of his 
business on a little black box with 
your name on it that he's going to plug 
into the wall, we better really think 
twice about service. And, that's one 
of the biggest areas that we find our 
energies going into. 

There are some nice things that 
micro-computers have going for them. 
One of the approaches we use is that 
instead of taking a service contract, 
which is expensive relative to purchase 
price (12$ of the purchase price of the 
equipment, per year) we say a micro- 
computer is so small you can just tuck 
it under your arm and throw it into the 
back of your car after work and bring 
it on down to the store and we'll fix 
it while you wait. Which is what we 
do. You can't do that very well with a 
System 32 or a Vydek word processor. 
So, there are some distinct advantages 
to being in the micro end of this but 
service is going to remain a big 

Support . Another one is support. I 
don't care how "canned" a package is, 
the phone rings continuously for the 
first three weeks that the system is 
sitting in a businessman's operation. 
You have to have people on hand who can 
answer those questions, who can find 
bugs in programs if they exist. That 
one's a real sleeper. Having enough 
people on hand all the time who can 
answer those phone calls and straighten 
things out and can also modify packages 
as they need to be modified. 

Software Tools . The third area is 
software tools. All the little things 
that the big guys have that we're just 
starting to develop. Things like 
utilities for sorting and data entry 



BOX 1579, PALO ALTO CA 94302 

and formatting. This is another area 
where Computer Power & Light has been 
working. Writing machine language sorts 
for the Micropolis disk BASIC and 
writing machine language data entry 
packages that prevent you from entering 
a four digit zip code. Things that 
prevent you from entering a part number 
using digits when it's expecting 
alpha-numeric characters. Things like 
that. It's a big area because when 
you're out in the business world, you 
know, the secretary might have been 
hired last week and doesn't know a 
computer from a Datsun . 

Training . The fourth area that we have 
always been strong in, and that we feel 
is important, maybe others may not 
agree... is training. We say that our 
training is better than Xerox's, and 
we've got the affidavits to prove it. 
It's an area in which we had some 
experience. The people who are part of 
Computer Power & Light who were 
successful teachers earlier in their 
careers have expended great effort in 
designing training sessions and courses 
that are effective. Our classes are 
not an afterthought, as they are with 
so many other computer companies. 
They've taken a lot of our resources, 
but we feel that they give us a pretty 
good advantage in dealing with the 
problems and remaining profitable in 
this industry. 

I hope my remarks have given each of 
you some insight into the realities of 
the low- cost business computer 
"revolution". I'll gladly answer any 
question you may have. Thank you. 

* * * * 



BOX 1579, PALO ALTO CA 94302 

Thomas P. Bun, MBA, MSEE, 2171 Sharon Road, Menlc Park, CA 94025 


MICROLEDGER is a General Ledger system, re- 
duced to the absolute essentials. Written in 
'BASIC, it employs only two files, Chart of 
Accounts and Journal. It will run in 8 kilo- 
bytes of user memory. 

The package is based on a foolproof , step- 
by-step procedure, designed with the novice 
computer user in mind. Some familiarity with 
simple accounting practices is required. The 
documentation includes a ready starter Chart 
of Accounts for the small business user. 

After entering data, changes are easily 
made , both to the Chart and to the Journal . 
Even after posting, adjustments may be made 
promptly, both to the Profit and Loss State- 
ment and to the Balance Sheet. 

Decision problem for business accounting 

Small business, just like its large enter- 
prise counterpart, faces an early decision in 
setting up its accounting practice. Should it 
first deal with the most urgent parts of its 
detail procedures, like inventory, receivables, 
payables? Or should it in the first place, set 
up an overall framework for its entire accoun- 
ting? This latter alternative corresponds to 
the proper organization of the general ledger 
as the first step. The detail procedures, then, 
follow as mosaic stones in a picture, whose 
major outlines already have been properly de- 

The purpose of MICROLEDGER is to supply 
this overall framework for the small business 
accounting. To choose this alternative is, 
therefore, equivalent to the TOP-DOWN systems 
approach, as opposed to the BOTTOM- UP approach. 
In this way, it is easier to design procedures 
that do dovetail together. The final goal, a 
totally automated comprehensive system, is made 
easier to achieve, since it can now be attained 
step by step. When a new module is set up, no 
change has to be made in old modules, since the 
general framework has been set up first. 

General Ledger packages 

There were hundreds of excellent general 
ledger packages available in the market. The 
need for a greatly simplified new package, like 
MICROLEDGER, arose when inexpensive microcom- 
puters suddenly became available to thousands 
of small businesses that could not afford to 

use computers before. The existing general 
ledger packages were designed for more elabo- 
rate machines and for users with more resour- 
ces. Typically, their use required people with 
computer background and training; not unfre- 
quently, a package would require 15 or 16 files 
to be set up and manipulated. Complex procedu- 
res and lengthy operating manuals were invol- 
ved. Most of todays microcomputer users would 
simply keep away from them, because they are 
too difficult to understand, set up and run.- 

Organization of MICROLEDGER 

The new package is set up in the form of six 
'BASIC programs. According to the particular 
hardware - and corresponding 'BASIC version - 
the programs may be stand-alone or chained. 
Each program is intensely interactive, main- 
taining a "conversation" with the user and 
leading him to the next step. 

LEDGER1 is the input program for the Chart 
of Accounts file. This is the general ledger 
master file. 

LEDGER2 is the. input program- for "the Journal 
file. This is the transactions file of the 
system. See an example, Fig. 1 in the Appendix. 

LEDGER3 lists both files, to help the user in 
checking them out and to supply audit trails 
for a complete story of all transactions. 

LEDGER4 runs the trial balance, displays the 
result in a table and then allows the user to 
ask for posting the 'transactions to the led- 
ger, Fig. 2, top part, in the Appendix. 

LEDGER5 reports the Profit and Loss Statement 
for the period covered by the journal. An 
example is shown on Fig. 2, bottom part. 

LEDGER6 reports the Balance Sheet at the 
closing date. A balance sheet is shown in 
Fig. 5 of the Appendix. 

A dual drive diskette device is recommended, 
in order to enable the user to keep the six 
programs on one drive, on a protected dis- 
kette, while the two data files, the Chart 
of Accounts file and the Journal file, are 
kept on another diskette on the other drive. 



BOX 1579, PALO ALTO CA 94302 


The price to pay for such a greatly- 
simplified setup is relatively small. The 
account numbers are kept to 3 digits and 
must obey to a predetermined structure: 

101-199 Current Assets 

201-299 Non-Current Assets 

301-399 Current Liabilities 

401-499 Long-Term Liabilities 

501-599 Owners Equities 

600 Retained Earnings 

701-799 Revenue Accounts 

801-899 Direct Expense Accounts 

901-999 General and Administrative 
Expense Accounts 

Fig. 4 

The number of journal entries is only 
limited by the capacity of the diskette. A 
typical device, like a Micropolis floppy 
diskette, could contain 10,000 transaction 
entries. This is much more than would be 
actually required and recommended for use 
in practice for a single posting. 

The final restriction is the fact 
that the system does not automatically 
handle a multiple department situation. 


MICROLEDGER was set up to allow for 
easy bypassing of this restriction. If a 
user would like to handle, a milt iple depart- 
ment situation, he can set up his Chart of 
Accounts for this, by using up to 10 acc- 
ounts for each purpose allowing for up to 
10 departments. 

Each department will have its own 
journal file. These will be kept on sepa- 
rate diskettes, with the same filename, 
but a physical label differentiating each 

At run time, the program asks for the 
name - now the department becomes identi- 
fied on the report for free. And the proper 
posting reference being employed, the led- 
ger now handles the multiple department 
situation correctly. 

Other extent ions will be added by 
COMPUMAX ASSOCIATES, the owner of the 
system, by the development of additional 
modules. The Accounts Receivable and the 
Accounts Payable modules are being de- 
veloped currently. These future packages 
can be run both as stand-alone applica- 
tions and as preprocessors to MICROLEDGER 
since their output files are compatibe in 
format with the MICROLEDGER journal file. 


Many small business users may set up this 
package with the sample Chart of Accounts as 
supplied. The sample chart is condensed from 
many actual small business users charts. This 
alone may, frequently, justify its employment. 

The way the chart is set up, it is very 
easy to make small changes, additions and de- 
letions for coustomization to any particular 
requirement . 

The balancing outputs reported at posting, 
and at printing the statements, help a great 
deal in detecting inconsistencies in the actu- 
al figures entered. 

Once the necessary adjustments are deter- 
mined by the user, these can be effected very 
easily, by the use of the update options in 
the programs. These allow for insertion, change 
and deletion of any data item. 

After the printout is finally acceptable 
to the user, it becomes an ideal frontispiece 
for a month of actual records. The printout is 
solid, clear, readable. It will be welcome by 
tax inspectors, I. R.S. auditors, accountants, 
and - particularly - by the user himself. 

The employment of consistent and perma- 
nent accounting practices, month after month, 
take away the drudgery from the bookkeeping 
chore. Instead, it becomes a pleasant and 
rewarding occupation, a useful tool for the 
.decision making by the. small business owner. 


Two premises were employed in the design 

- keep it as simple as possible, in con- 
trast to extisting general ledger 
packages with too many options, files 
and complications; 

- make it, nevertheless, powerful enough 
to accomodate the needs of a general 
small business user community. 

It is obvious that many tradeoffs had to 
be made to accomodate these premises, that 
so often lead to conflicting requirements. 

It is the hope of the author of this 
paper, that the product described met both 
these goals satisfactorily. 

It is a pleasant duty to offer my best 
thanks for the significant assistance received 
during the development of this product, from 
the persons listed under ACKNOWLEDGMENTS. 



BOX 1579, PALO ALTO CA 94302 






2 07-01-77 J . STERN PAYMENT 


4 07-04-77 CHECK * 226 U.C.B. 

5 07-01-77 J.EUING NOTE 

6 07-17-77 E. P. R.I. CONTRACT 



9 07-31-77 PARTS SENT TO SHOP 



12 07-22-77 NEW ROOF ON SHED 

13 07-23-77 NOVA 3 COMPUTER 




17 07-31-77 'JULY SALES 



20 07-31-77 PRODUCTION 

21 07-31-77 STOCKROOM 

22 07-31-77 SUPPLIES 

23 07-31-77 JULY PAYROLL 

24 07-31-77 SOCIAL SECURITY 



27 07-31-77 BILLS PAYABLE 


29 07-31-77 RENTALS PAID 



































































BOX 1579, PALO ALTO CA 94302 

DATE OF 7 - 31 - 77 











< 1215 ) 









( 1845 



( 6808 



( 12840 



( 21455 



( 2050 























Fie. 2 



BOX 1579, PALO ALTO CA 94302 













TO R. E. 




































Fie. 3 





BOX 1579, PALO ALTO CA 94302 



The following textbooks were used during 
the development of MICROLEDGER: 

Earl A. Spiller: "Financial Accounting" 
Irwin, 1971 

Richard Mattesich"Accounting and Analy- 
tical Methods", Irwin, 1964 

John Dearden Hj F.Warren McFarlan: "Manage- 
ment Information Ststems", Irwin, 1966 

Anthony, Dearden and Vancil: "Management 
Control Systems", Irwin, 1965 

These books might be found at the libra- 
ries of the University of Santa Clara 
and Stanford Business Schools. 

The following persons were of significant 
assistance during the development of the 

Paul J. Terrell, as President of Byte, Inc. 
served as inspiration and support from the 

Mike D. Lipschutz and John K. Borders of 
Microcomputer Software Associates - MSA, 
contributed with valuable advice. 

Boyd Wilson and Mike Black, of the Mountain 
View Byte Shop, contributed with many hours 
of patient help in setting up the hardware. 

Zulmira McMorrow of the McMorrow Engineering 
Group, one of the first users, completed the 
first large-scale data entry. 


Thomas P. Bun 

CompuMax Associates 
505 Hamilton Avenue 
Palo Alto, CA 94301 

Thomas P. Bun was born in Budapest, 
Hungary. Having graduated in general 
engineering, with a major in planning, 
from the Budapest Polytechnical Univer- 
sity, he worked in Europe and Latin Ame- 
rica, finally settling in California. 
Mr. Bun obtained an M.S. degree from Stan- 
ford University, majoring in Digital Sys- 
tems/Computer Science, and an M.B.A. 
from the University of Santa Clara. 

Positions occupied by Mr. Bun have included: 
senior systems analyst with Stanford Research 
Institute; project manager in microcomputer 
application design with Rockwell Internatio- 
nal; manager- systems with Light S.A., the 
largest privately owned electric power uti- 
lity in Latin America; director of Reduto, 
a Brazilian engineering company; president 

of.. ..the Latin American Astronomical League; . 

member of the Council of the National Com- 
mission for Space Activities of Brazil; 
management science analyst with I. S.I. Cor- 
poration and product engineer with Smith- 
Corona Marchant Corporation. Since January 
1977, Mr. Bun is president of CompuMax 
Associates, a Palo Alto-based software 
house, specializing in the production of 
application packages for microcomputers. 



BOX 1579, PALO ALTO CA 94302 

Money for Your Business-Where to 
Find It, How to Get It 

Don Dible 

468 Robert Rd. 

Life Insurance A loan based on the cash value of your life 
insurant policies can be a low-interest source of money Many 
policies provide for automatic loans from the insurance earner at 
interest rates far below the prime rate charged by commercial 

Vacaville CA 95688 banks. 

One of the recurrent problems that plagues most owner- 
managed businesses is the shortage of adequate capital. In the case 
of new businesses, this problem can be severe. 

Promising small businesses traditionally have relied upon such 
sophisticated money sources as investment bankers, venture 
capitalists,, and federally licensed and leveraged Small Business 
Investment Companies (SBICs) for their equity capital needs. In 
today's market, these sophisticated investors have many options 
from which to choose. Regardless of market conditions, their 
objective continues to be the maximization of return on invest- 
ment, consistent with intelligent risk analysis. 

Their first option is'the purchase of securities in publicly traded 
companies at bargain basement rates. Their second option is the 
private placement of growth capital in young, but already estab- 
lished, companies of demonstrable merit. In a tight money market, 
the terms of such an arrangement may be very attractive for the 
investor. The investment option representing the highest risk, 
obviously, is in financing the start-up company. Given the first 
two options, a private placement with a start-up company is, I 
think you'd agree, a most unlikely choice as an investment vehicle. 

What about banks? Surely, you can get a personally endorsed 
loan for your business from the friendly loan officer at your 
commercial bank. Not necessarily. In evaluating the financial state- 
ment of any business, a banker will pay particular attention to the 
ratio of debt to equity capital. A general rule is that you cannot 
borrow more money than has been invested as equity capital. 
Furthermore, most bankers are not eager to lend money to new 
businesses at all, preferring to wait until there is at least some 
history of profitable operation. 

How, then, can you assemble the resources you'll need to get 
started and to keep you going until you begin to turn a profit? 
Take heart; there is hope. There are, in fact, four different money 
sources that can now help you stretch your business dollars: 1) 
personal financial resources, 2) trade credit, 3) customers, 4) 

Personal Financial Resources 

Many of us significantly underestimate the value of our personal 
resources when taking a financial inventory. Let's look at some of 
the more obvious (and not-so-obvious) resources you may have: 

Home In the recent inflationary period, the value of practically 
all residential real estate has appreciated substantially. If you have 
owned your own home during this period, you are indeed fortun- 
ate. The difference between the current value of your home and 
the balance outstanding on your mortgage may represent an 
equity asset of between $20,000 and $50,000. You might consider 
converting this asset to cash by 1) taking a second mortgage in the 
amount of your equity, 2) refinancing your mortgage, 3) subdivid- 
ing your lot and selling part of it, or 4) selling your house and 
moving into an apartment. Every year tens of thousands of people 
go into business using "house money" to bankroll their ventures. 


Stocks and Bonds You may own stocks and bonds that, for a 
variety of reasons, you do not wish to sell. Such instruments make 
ideal collateral for loans, and you can borrow anywhere from 40% 
to 60% of their current value, depending on the lender. 

Credit Cards and Personal Credit When you prepare vour finan- 
cial business plans, you will surely want to take the fullest'possible 
advantage of your credit worthiness. At the same time, you will 
want to minimize the amount of money that you have to take out 
of the business to meet your personal financial needs. This is 
particularly true when the business is just getting started and every 
dollar in the treasury is needed to finance business growth. During 
this period, personal credit cards of all types and descriptions can 
give you just the extra leverage you need. Instead of drawing a 
heavy salary in the first year of operation, supplement your 
modest cash draw with the judicious use of credit cards. 

Credit cards come in a great variety of classifications: bank 
cards such as Master Charge and BankAmericard, travel and enter- 
tainment (T & E) cards such as Diners Club and American Express, 
as well as cards for department stores, oil companies, airlines, and 
so on. With a good credit record, it should be no great feat for you 
to secure enough credit cards to give you a personal line of 
revolving credit in excess of $30,000. 

Now, obviously, you can't buy a carload of raw materials for 
your business with your bank credit card, but you can cover a lot 
of your personal expenses with it — as well as such business expen- 
ses as transportation, food, and lodging. With a little thought, I'm 
sure you can come up with a variety of ways to employ credit 
cards in financing your particular business. 

While we're on the subject of personal credit, you should give 
serious consideration to arranging for the installment purchase of 
an automobile and other major consumer items before you start 
your business. You'll find it a lot harder to qualify for this kind of 
financing after you've started. Lenders tend to worry a great deal 
about the financial stability of new small businesses and their 
founders. Your credit application looks a lot better when, under 
"Current Employer," you show that you have been employed 
with "Solid as a Rock, Inc." for the last six years instead of with 
"Shaky at Best" for the last six weeks. 

Relatives and Friends I happen to consider relatives and friends 
to be extremely valuable personal financial resources. Where else 
can you find lenders who, simply because they like you, will 
advance money on anything as risky as starting a new business? 
However, do yourself and your lenders a favor. Document their 
loans. Draw up a formal note showing interest charged and the 
dates on which principal and interest are payable. You never know 
when a personality problem may arise that could precipitate 
calling the loan. That can become awfully messy. 

Take the case of the young man whose rich aunt loaned him 
$100,000 with which to start his business. The formality of a loan 
agreement was ignored. Six months later the aunt died. The heirs 
then forced the entrepreneur to liquidate his business so that they 

267 BOX 1 579, PALO ALTO CA 94302 

could each get their share of the aunt's estate. 

Relatives and friends can also come in handy as cosigners. Let's 
assume that you apply for a loan and the lender decides that your 
qualifications are marginal. The availability of a financially strong 
cosigner can swing the balance in your favor. Your relatives and 
friends can prove to be very important personal assets. Don't 
overlook them in your financial planning. 

Trade Credit 

According to a U.S. government study, trade credit constitutes a 
33% larger factor in financing the business community than 
that represented by bank loans. Clearly, every company treasurer 
should give serious consideration to trade credit in financing a 
small business. 

Customarily, suppliers provide trade credit as an inducement to 
their customers to do business with them. Although credit terms 
in most industries are extended on a net 30-day basis, overall 
averages may stretch this to 45 days, 60 days, or even longer 
depending on the condition of the overall economy and the 
particular industry involved. Under the guise of "cash manage- 
ment," many companies make a habit of vigorously enforcing 
their collection policies while simultaneously treating their 
accounts payable with cavalier disregard. 

Just how you handle your payables is your own business. But 
you should be aware of the fact that not everybody in business 
makes a habit of paying bills the day the first invoice arrives. 

When it comes to making use of trade credit as a tool in 
financing your business, I recommend that you establish with your 
suppliers, in advance of purchase, the best extended credit terms 
you can negotiate. Depending on how hungry your suppliers are 
for your business, you may be able to arrange extremely attractive 
terms. Securing competitive bids will greatly improve your bar- 
gaining position. However, once you have agreed to terms of 
payment, honor your commitment rf you value your credit. 
Nothing is more difficult than trying to operate a business when 
your suppliers will ship to you only on a COD basis. 


There are many ways in which customers can be induced to help 
you finance your business. The degree to which they are vHlling to 
do so depends on the extent to which you enjoy a monopoly on 
the goods or services that you offer. In other words, no one will 
pay in advance if he can get essentially the same goods or services 
from another supplier under more liberal credit terms. However, if 
you have the only game in town, you may be in a position to get 
your customers to finance your business. 

As you know, the telephone company, electric and gas com- 
panies, the post office, and a host of government monopolies 
require deposits in advance (independent of your credit- 
worthiness), so that you may enjoy the benefits of their services. 
The same kind of policy may be applied in the operation of 
certain small businesses and selected industries. 

A San Francisco Peninsula electronics company, which sells 
a patented device available nowhere else, provides a good illus- 
tration of customer financing. The company's customers are 
required to make a deposit of one-third of the purchase price 


when the order is placed. Another one-third is payable on delivery, 
and the balance is due in 30 days. As a result of these favorable 
terms of sale, this company has been able to finance a highly 
satisfactory growth rate while experiencing almost no problems 
with cash flow. 

An extreme example of this type of financing is the case in 
which the customer is required to pay the full purchase price at 
the time the order is placed. You may be surprised to learn that 
you have been financing certain of your own personal suppliers on 
this basis for years. I am referring, of course, to magazine publish- 
ers. If you take a three-year subscription to a magazine, you may 
get the first one or two issues on credit based on your promise to 
pay. However, in order to continue receiving the magazine, you 
must pay the subscription bill, even though you won't receive 
your final shipment for almost three years. 

Hugh M. Hefner, publisher of Playboy magazine, played this 
customer-finance game with admirable success when he initially 
offered lifetime subscriptions for $100. Obviously, he needed the 
money to finance his growth. Had the subscribers to some of the 
early issues bought $100 worth of Playboy stock instead of a 
subscription, they could have picked up a tidy profit. 

In many industries, including a number of the construction 
trades, it is usual for the suppliers to receive progress payments 
when previously agreed-upon levels of project completion are 
achieved. Highway construction, aircraft assembly, and other 
large-scale projects are often financed in this way. 

A special form of customer financing occurs in a number of 
service and manufacturing industries. Here the customer provides 
raw materials that the vendor transforms into finished goods. In 
the book printing trade, a publisher may supply the printer with 
paper; in a machine shop, the customer may provide the metal to 
be worked; and in a tailoring service, the customer may provide 
the cloth to be fashioned into a dress or a suit. In each case, the 
vendor avoids the expense of financing the purchase of raw mate- 
rials.- - 

Another way of getting the customer to pay in advance for 
goods or services is to establish some kind of "membership" 
arrangement. Here the customer may pay an annual fee for the 
privilege of attending meetings. Using a similar membership 
approach, some discount department stores require customers to 
pay a membership fee for the privilege of shopping there. 

A novel twist on this customer-backed approach to business 
finance is seen in physical fitness spas, dance studios, and other 
contract service organizations. Customers sign an agreement to 
purchase the service offered for a period of one or more years. In 
most cases, the intentions of the customer are honorable and 
sincere at the time he executes the contract. However, many 
people lose interest in fitness and other self-improvement pro- 
grams after a short time, and the incentive to pay the installments 
on the service contract may falter. In anticipation of these long- 
range collection problems, the original holders sell the contracts to 
finance companies at a substantial discount. While the average 
individual may balk at paying the original contractor for services 
not used, he is more likely to pay a finance company when 
notified that it has taken over his contract. The result is that the 
original contractor gets a handsome chunk of cash almost immedi- 
ately after the customer signs the contract, whether or not the 
customer continues to use his facilities. Once again, the customer 
has financed the business, albeit indirectly. 

268 BOX 1 579, PALO ALTO CA 94302 

I Economizing 

I In preparing your pro forma financial statements, you probably 
I allocated quite a bit of money for office equipment and other 
I capital expenditures. If you figured on buying new equipment, 
I figure again. Used equipment is what you want. That way, you 
■ may find that you need a lot less cash than you originally thought. 
Forget the rosewood paneled office with the Italian marble-topped 
desk, too; that comes later. And most financially strapped entre- 
preneurs quickly learn the delights of night coach and special tour 
package rates to save a few dollars when traveling. There are many 
other ways to economize in the operation of your business; we'll 
get to them shortly, but first let's concentrate on where to find 
used equipment. 

1. Used equipment dealers may handle anything from office equipment 
to laboratory test equipment to cash registers to display cases for 
butcher shops and retail stores. You'll find these dealers listed in the 
Yellow Pages of your telephone directory under such headings as 
"Used Equipment Dealers," "Second Hand Dealers," and "Suiplus 
Merchandise." You'll also find dealers listed under generic headings 
such as "Office Furniture— Used." 

2. Dealers in new merchandise invariably find themselves stuck with a 
variety of goods that cannot truly be represented as new. "Demon- 
strator" and "loaner" equipment (provided for the temporary use of 
customers who are awaiting delivery of new equipment or who arc 
having their own equipment repaired) fall into this category. These 
dealers may also have floor samples, warehouse- and freight-damaged 
goods, and obsolete-but-serviceable rental equipment available at a 
substantial savings. 

3. Bankruptcy and liquidation auctions provide an opportunity to get 
some real bargains. To secure information on where and when 
auctions are to be held, consult the Yellow Pages of your telephone 
directory under "Auctions" or "Auctioneers." Many auctioneers 
maintain mailing lists of interested clients and send out brochures 
announcing forthcoming auctions. 

You may also obtain information on bankruptcy auctions by 
contacting the bankruptcy court in your area. You'll find this court 
listed under "U.S. Government" in the White Pages of your tele- 
phone directory. 

4. Bankrupt companies that receive protection under Chapter 1 1 of the 
Bankruptcy Act are very likely to be interested in liquidating some 
of their assets. If you are lucky enough to learn about such a 
company in your own industry, you are in a unique position to fill 
your equipment needs quite reasonably. Bankruptcy filings are 
announced in the legal newspapers of record serving various com- 
munities across the country. Also, when a large company in a 
particular industry files for voluntary bankruptcy under Chapter 1 1, 
the trade journals serving that industry will usually carry mention of 
this fact. Upon learning of such a bankruptcy, don't hesitate to call 
the company involved to determine whether the owners are inter- 
ested in selling some of their assets. 

5. Now and then major corporations simply decide to get out of a 
particular industry and shut down one of their divisions. A friend of 
mine who has his own company read about such an instance in a 
trade journal. He purchased a $100,000 (price when new) piece of 
test equipment for a mere $7,000 cash. Then he called a leasing 
company and received $45,000 for the same equipment when he 
agreed to lease it back from them over a five-year period. He realized 
an immediate cash infusion of $38,000. 

6. The United States Government is the single largest consumer of 
goods and services in the world. Not surprisingly, it also disposes of 
enormous quantities of surplus goods on a more or less continuous 
basis. For information on the sale of surplus government equipment 
at locations all over the world, write to the Department of Defense 
Surplus Sales Office, Box 1370, Battle Creek, Michigan 49016, and 
the Assistant Commissioner for Personal Property Disposal, Federal 
Supply Service, General Services Admin